Living in a fungal world: impact of fungi on soil bacterial niche development⋆

Abstract

The colonization of land by plants appears to have coincided with the appearance of mycorrhiza-like fungi. Over evolutionary time, fungi have maintained their prominent role in the formation of mycorrhizal associations. In addition, however, they have been able to occupy other terrestrial niches of which the decomposition of recalcitrant organic matter is perhaps the most remarkable. This implies that, in contrast to that of aquatic organic matter decomposition, bacteria have not been able to monopolize decomposition processes in terrestrial ecosystems. The emergence of fungi in terrestrial ecosystems must have had a strong impact on the evolution of terrestrial bacteria. On the one hand, potential decomposition niches, e.g. lignin degradation, have been lost for bacteria, whereas on the other hand the presence of fungi has itself created new bacterial niches. Confrontation between bacteria and fungi is ongoing, and from studying contemporary interactions, we can learn about the impact that fungi presently have, and have had in the past, on the ecology and evolution of terrestrial bacteria. In the first part of this review, the focus is on niche differentiation between soil bacteria and fungi involved in the decomposition of plant-derived organic matter. Bacteria and fungi are seen to compete for simple plant-derived substrates and have developed antagonistic strategies. For more recalcitrant organic substrates, e.g. cellulose and lignin, both competitive and mutualistic strategies appear to have evolved. In the second part of the review, bacterial niches with respect to the utilization of fungal-derived substrates are considered. Here, several lines of development can be recognized, ranging from mutualistic exudate-consuming bacteria that are associated with fungal surfaces to endosymbiotic and mycophagous bacteria. In some cases, there are indications of fungal specific selection in fungus-associated bacteria, and possible mechanisms for such selection are discussed.

Fungal–bacterial interactions, Fungus-associated bacteria, Competition, Mutualism, Mycophagy

1 Introduction

Most of life's history has been in the aquatic environment. Here, in the past, bacteria have ruled the two most important processes in organic matter cycling, namely primary production (autotrophy) and decomposition (heterotrophy). Nowadays, though eukaryotes, in particular algae, are major contributors to primary production, bacteria are still by far the dominant decomposers in water and sediments [1].

From the late Ordovician to the early Devonian on, however, land began to be colonized with plants, being the majority autotrophs. This was probably also the time when soil development started [2–4]. Fossils of the first land plants have revealed that filamentous fungi, resembling the glomalean mycorrhizal fungi, were present in their root tissues [4, 5]. In fact, it has been hypothesized that the development of plant-fungal mutualism preceded the development of roots and, therefore, has been crucial to the ability of plants to colonize the land [4]. The mycorrhizal association between fungi and roots of plants has remained very successful through evolution [6]. Almost all successful extant land plants (?80% angiosperms, 100% gymnosperms and 70% pteridophytes) in nature are associated with one or several mycorrhizal fungi. Mycorrhizal formation has not been restricted to the Glomales but is also particular in Basidiomycota. In most cases, the basis of the mutualism is that the plant provides the major source of fixed carbon, whereas the fungus provides the host with mineral nutrients, water and in some cases protection from root pathogens.

The colonization of land by plants provided a new habitat for heterotrophs. The soil habitat is essentially different from that of sediments in that it contains air-filled voids. The inability of the bacterial unicellular body form to bridge these air-filled voids restricts bacterial motility in soils. This disadvantage is overcome by the hyphal/mycelial growth form [7]. Furthermore, given their ability to translocate nutrients [8], fungal hyphae are better adapted to cross nutrient-poor spots in soil than bacteria in searching for the heterogeneously distributed nutrient resources.

In addition to differences in three-dimensional space between soils and aquatic environments, there are also strong differences in the composition of the organic matter that they receive [9]. On land, plants had to protect themselves against desiccation and radiation, and they developed new structural compounds, e.g. cutin and suberin that cover leaves, stems and roots [9]. This development provided the terrestrial decomposer microorganisms with a new challenge, i.e. gaining access to the well-protected organic matter. Hence, efficient decomposition of bulky plant tissues requires penetration – a feature of the hyphal growth form.

The clear advantages of the hyphal growth form to colonize soil and to penetrate vascular plant tissues is probably the reason that the primitive mycorrhiza-like fungi could evolve to occupy decomposer niches. Both fossils and molecular biological data indicate that major groups of saprotrophic soil fungi, Zygomycota, Ascomycota and Basidiomycota, are descendants of the glomalean fungi [2, 10, 11]. Interestingly, an important terrestrial group of bacterial decomposers, the actinomycetes, have also developed the hyphal growth form [7], but this has not enabled bacteria to monopolize terrestrial heterotrophic processes.

The evolution of fungal decomposers not only included optimization of colonization and penetration abilities, but also the development of new pathways to degrade recalcitrant structural compounds that are unique to vascular plants [12–14]. This is especially the case for lignin. With cellulose fibrils embedded in a lignin matrix, plants developed a flexible but very robust structural framework. This lignocellulose complex is both physically and chemically resistant to degradation even after death of the plants. The evolution of pathways for lignin degradation has been largely restricted to Basidiomycota and xylariaceous Ascomycota [15, 16].

With the monopolisation of two important niches in terrestrial ecosystems, namely mycorrhiza formation and decomposition of lignocellulose, fungi have obtained a pivotal role in the functioning of terrestrial ecosystems. This must have had a strong impact on the evolution of terrestrial bacteria. Bacteria and fungi have undergone niche differentiation in the decomposition of organic material. The evolution of this differentiation is an ongoing process. In this review, we will consider this aspect with respect to the decomposition of plant-derived organic matter, including easily assimilable root exudates (Section 2.1), cellulose (2.2) and recalcitrant lignin (2.3). Some bacteria have evolved the potential to utilize products released from complex organic substrates as a result of fungal exoenzyme activity (2.4). The presence of fungi, however, has resulted not only in the loss of potential niches for bacteria, but also in the creation of new ones. These are discussed in the second part of this review, which addresses bacteria growing on fungal exudates (3.1, 3.3), living hyphal compartments (3.4 and 3.5) and the walls of dead hyphae (3.6).

So far, soil bacteriologists and mycologists have largely neglected each other's research fields [17]. We hope that this article will inspire them to an intensive collaboration.

2 Niche differentiation between bacteria and fungi with respect to the decomposition of plant-derived substrates

During the evolution of terrestrial microbial life, fungi have become the major decomposers of recalcitrant organic matter. Bacteria on the other hand have been able to maintain a significant role in the degradation of simple substrates. However, this is only the general picture and, as it will be shown below, there is an ongoing confrontation between fungi and bacteria for both complex and simple substrates.

2.1 Role of fungi and bacteria as decomposers of simple substrates: root exudates

Plant roots exude substantial amounts of low molecular weight organic compounds such as amino acids, sugars and organic acids, resulting in increased microbial populations and activity [18–21]. Given the high numbers of fast-growing bacteria, especially Gram-negative strains, in the zone close to the root (rhizosphere), it has been widely assumed that easily degradable plant exudates are almost exclusively degraded by bacteria [18, 20]. Therefore, most publications on dynamics of microorganisms in the rhizosphere deal only with shifts in bacterial communities. However, saprotrophic and plant pathogenic fungi have been isolated frequently from the rhizosphere and rhizoplane (root surface) [22–24]. With respect to the saprotrophic fungi, species with ruderal characteristics, including rapid growth, prolific spore production and ability to use only relatively simple fixed carbon compounds, appear to be rhizosphere inhabitants [22, 23, 25]. This is true, for example, for members of the zygomycetous order Mucorales.

Until recently, it has been difficult to estimate the relative contribution of fungi and bacteria to decomposition in localized environments such as the rhizosphere. Most studies have relied on selective respiration, i.e. the respiration that occurs after addition of selective (fungal or bacterial) antibiotics [26]. Although such studies have indicated the potential for fungi to compete for simple substrates, they do not provide a reliable estimate of the competitive strength of fungi versus bacteria, because of artefacts due to mixing, severing of hyphae, non-target inhibition, incomplete inhibition and high substrate loading. The advances in analytical techniques that allow measurement of stable isotopes (¹³C) incorporated in biomarkers (e.g., signature fungal and bacterial phospholipid fatty acids or the characteristic fungal membrane sterol ergosterol) have created the possibility of a greatly improved evaluation of the relative importance of fungal and bacterial decomposition [27].

Recent studies using ¹³C-labelled grasses indicated that fungi might have a significant contribution to the decomposition of root exudates [28, 29]. Similarly, incorporation of ¹³C into fungal phospholipid fatty acids was seen after addition of ¹³C-labelled model rhizosphere compounds to soil [30–33]. These studies have revealed that the fungal contribution to the decomposition of easily degradable substrates is highest in acid soils and at high substrate loading rates. This pattern has been attributed to the ability of fungi in its superior osmotic stress tolerance capabilities in comparison with those of the bacteria [31].

At low substrate concentrations, microorganisms with high-affinity uptake systems are expected to have the advantage. However, since such high-affinity systems have been described for both bacteria and fungi, low substrate concentrations do not necessarily confer an advantage on either group [7]. Clearly, there is a need for more studies that address the effects of substrate concentration and composition on the relative competitive strength of fungi and bacteria. In the near future, more detailed information on the composition of fungal and bacterial decomposer active in situ in the rhizosphere, and in other environments, will become available from substrate-induced labelling of DNA and RNA (stable isotope probing) [34].

The presence of filamentous fungi with the ability to rapidly decompose easily degradable organic compounds must exert a selection pressure on bacteria to compete for these nutrients. A multitude of antifungal strategies has been revealed in bacteria, including production of directly inhibitory factors such as HCN, antibiotics, lytic enzymes and volatiles (interference competition), as well as nutrient-sequestering factors such as iron-chelating siderophores (substrate competition) [35–38]. The nature and regulation of production of such antifungal compounds has been intensively studied with the aim of improving the possibilities for biocontrol of root-infecting and pathogenic fungi [39]. Several of the bacterial competitive strategies, however, appear to be important not only in the interaction with fungi but also in the interactions with other bacteria, e.g., in the rhizopshere. For instance, antibiotics produced by rhizosphere-inhabiting Pseudomonas spp. are active against a range of fungi and bacteria [40]. Hence, it is often not clear to what extent an antifungal trait has actually evolved as a strategy to compete with fungi. Experimental manipulation of fungal density could be used as an initial approach to study the impact of fungi on the selection of bacteria with anti-fungal properties.

The fungi themselves often produce compounds with antibacterial activity, and have developed several strategies to counteract bacterial antagonism including detoxification, removal of antibiotics by efflux, and modification of bacterial gene expression [41]. Hence, it appears that the bacterial and fungal inhabitants of the rhizosphere are involved in an ongoing evolutionary arms race.

2.2 Role of fungi and bacteria as decomposers of complex substrates: cellulose

Cellulose is the most abundant organic compound on the planet (30–50% of plant dry weight) and, therefore, represents a huge source of energy for microorganisms. Microbial decomposition of cellulose can take place under both aerobic and anaerobic conditions. The ability to degrade cellulose aerobically is widespread among fungi and is especially well represented among the Ascomycota and Basidiomycota (groups now to considered to encompass the corresponding asexually reproductive states formerly grouped separately as Deuteromycota) [15, 42, 43]. Aerobic cellulose degradation is also known for several soil bacterial species in both filamentous (e.g. Streptomyces, Micromonospora) and non-filamentous (e.g. Bacillus, Cellulomonas, Cytophaga) genera [43].

Fermentative degradation of cellulose is widely distributed among obligately anaerobic bacteria (e.g. Acetivibrio, Clostridium, Ruminococcus) but is also present among some fungal species, belonging to the phylum Chytridiomycota, that inhabit the gastrointestinal tracts of ruminants [43]. However, in most anoxic environments, bacteria are almost exclusively responsible for degradation of cellulose [43, 44]. The anaerobic bacteria use complexed cellulase systems, cellulosomes, which allow concerted enzyme activity of different cellulases, and minimize the distance over which the cellulose hydrolysis products must diffuse to the cells [43, 45]. Remarkably, this highly efficient cellulolytic system is not present in aerobic bacteria. Aerobic cellulolytic fungi and bacteria produce freely diffusible extracellular cellulase enzyme systems consisting of endoglucanases, exoglucanases and β-glucosidases that act synergistically in the conversion of cellulose to glucose [43, 46]. The cellulases are distributed over different enzyme families suggesting convergent evolution towards the same substrate specificity [43]. Thus, although components of the cellulolytic system of soil bacteria and fungi can be distantly related, their functionality is quite similar.

The presence of functionally equivalent cellulolytic systems in bacteria and fungi implies that competition for cellulose may take place in soil. In general, however, it is assumed that by far the most of cellulose degradation in soils is performed by fungi. This is probably due to the relative inaccessibility and to the chemical nature of cellulose. Cellulose fibres are seldom found in pure form, but embedded in a matrix of other structural polymers, in particular hemicelluloses and lignin. Furthermore, most cellulose in plant cell walls is resistant to enzymatic hydrolysis because of the crystalline (fibrillose) structure. Variation in the degree of accessibility and crystallinity of cellulose may lie at the basis of niche differentiation among different cellulolytic microorganisms in different substrates, such as wood or compost.

The hyphal growth of fungi and cellulolytic actinomycetes appears to be an important strategy to access cellulose fibres, via pores in the cell wall material, and to bring the cellulases into close contact with the cellulose polymers [43]. In lignified plant material, those fungi that cause ‘white rot’ are able to gain access to the cellulose within the lignocellulose complex by decomposition of the lignin polymers [43, 47, 48]. However, the ability to decompose lignin is not widespread, and some fungi, such as those causing brown rot or soft rot, access cellulose in woody tissues by other means (Table 1, Section 2.3). The fungi that cause soft rot (Type 1), produce fine penetration hyphae that give them access to the S₂ woody cell wall layer, in which they form chains of diamond-shaped cavities in the immediate vicinity of the hyphae, that follow the orientation of the cellulose microfibrils [15, 49]. The fungi that cause ‘brown rot’ are able to modify lignin, but they decompose it only to a limited extent. They decompose crystalline cellulose through the concerted action of enzymes (mostly endoacting, e.g. endoglucanases) and non-enzymatic systems [17, 50, 51]. The latter may include reduction of pH (e.g., by excretion of oxalate) and the excretion of iron containing low molecular weight glycopeptides that generate hydrogen peroxide. Free radicals are then produced (by the Fenton reaction: Fe²⁺+ H₂O₂? Fe³⁺+ OH ⁻+ OH⁻), which presumably diffuse freely into the S₂ layer of woody cell walls and eventually cause the depolymerisation of cellulose. A range of hypotheses has been put forward concerning how low molecular weight metabolites, metals and radicals function together in the degradation process, which has been reviewed elsewhere [51].

1

Summary of the characteristics of the major types of wood decay

Decay type	Cell wall constituents used	Anatomical features	Causal agents (major groups)
White rot: simultaneous	All	Walls attacked progressively from lumen	Basidiomycota and some Ascomycota
White rot: sequential/selective	All, though hemicelluloses and lignin used selectively initially	Walls attacked progressively from lumen	Basidiomycota and some Ascomycota
Brown rot	All except lignin, but the latter is modified	Entire wall attacked	Basidiomycota and few Ascomycota
Soft rot: type 1	All except lignin, but the latter is modified	Longitudinal cavities develop in S2 wall layer	Ascomycota, Deuteromycota and some bacteria
Soft rot: type 2	All except lignin, but the latter is modified	Walls attacked from lumen (in conifers mainly the S2 layer)	Ascomycota, Deuteromycota and some bacteria

Microscopic observations indicate that filamentous actinomycetes are almost never involved in wood decay [49]. This is surprising, as these hypha-forming, substrate-penetrating bacteria might at first appear to be the most likely direct competitors of cellulolytic fungi. Many cellulolytic actinomycetes, however, are unable to degrade crystalline cellulose [52, 53]. Further, their cellulases and hemicellulases operate best at neutral to alkaline pH, in contrast to the fungal enzymes that perform best at low pH [54]. Hence, the often acidic nature of wood [15], and the increased acidification that generally accompanies fungal growth and activity in wood, may prevent growth of cellulolytic actinomycetes [47, 51]. Moreover, actinomycete-mediated decomposition of cellulose can be substantial in composts that remain alkaline due to high ammonification [52]. Thus, it appears that a high pH and pH-increasing processes, like ammonification, are important factors mediating the level of cellulose degradation brought about by actinomycetes. The fact that many streptomycetes are able to exude antifungal compounds may indicate that they are competitors of fungi under appropriate environmental conditions [55, 56].

Information on the extent of in situ cellulose decomposition by non-filamentous soil bacteria is very limited. It seems likely that they are confined to easily accessible cellulose, due to their restricted ability to penetrate solids. Upon addition of cellulose to an agricultural soil, an initial phase featuring predominantly bacterial cellulose decomposition was recognized, followed by a stage dominated by fungal cellulose decomposition [57]. This could point to an opportunistic strategy of cellulolytic soil bacteria whereby they respond immediately whenever easily accessible cellulose is present. It is to be expected that such a strategy would also involve the production of inhibitory compounds, to protect the cellulose resources they colonize from invasion by actinomycetes and fungi. However, although the potential to produce antifungal metabolites has been shown for some cellulose-degrading bacteria, e.g. Bacillus pumilis, this has not been studied in the context of cellulose degradation [58].

Non-filamentous cellulolytic bacteria are also sometimes present in decaying wood, though decay is usually limited and they are often restricted to parts of the wood containing easily accessible pectin, cellulose and hemicellulose, e.g. ray cells, and to wood under environmental conditions inimical to fungal growth [15, 49, 59, 60]. Two patterns have been discerned: tunneling (or cavitation) in the S₃ layer of the cell wall in the approximately outer first cm of wood, and erosion, characterised by depressions, channels and ‘honeycomb’ patterns [59].

Besides the non-filamentous bacteria that have the ability to degrade crystalline cellulose, many other soil bacteria appear to have incomplete cellulolytic systems [61]. It may be that the cellulases of these bacteria, in combination with other hydrolytic enzymes, participate in the penetration of living plants in either pathogenic or endophytic associations.

2.3 Role of fungi and bacteria as decomposers of complex substrates: lignin

Lignin was present in the oldest known land plants, and it appears that the structural rigidity that it imparts was a prerequisite in the colonization of the terrestrial environment [12, 62]. Despite its widespread distribution and early appearance in terrestrial life, decomposition of lignin is largely, but not exclusively, found in certain genera of Basidiomycota that are collectively named white-rot fungi (Table 1) [15, 16, 48, 63]. The breakdown of lignin is mediated by enzymes, such as laccases and peroxidases, and free radicals, and occurs strictly under aerobic conditions. Bacterial lignin degradation appears to be negligible in terrestrial environments compared to the activity of white-rot fungi [63, 64]. However, growth of both filamentous and non-filamentous bacteria on lignin-like compounds has been observed [65–68]. Several actinomycete species have been shown to solubilise lignin, in particular lignin in grasses [63, 69]. This may be a strategy, similar to that of brown-rot fungi (Section 2.2), to gain access to cellulose.

2.4 Possible interactions between bacteria and fungi in relation to breakdown of the lignocellulose complex

It is clear from the preceding paragraphs that, under aerobic conditions, lignocellulolytic substrates are mainly broken down by fungi, despite the ubiquity of bacteria. However, bacteria present in lignocellulose-rich substrates may interact with fungal degraders in several ways.

The hydrolytic activity of extracellular fungal enzymes in lignocellulose-rich material results in production of water-soluble sugars and phenolic compounds [65, 70, 71]. These organic solutes comprise the actual sources of energy and carbon for the fungus. They are, however, also good growth substrates for other microorganisms, and can be competed for by bacteria. Intense bacterial competition could deprive the fungus of its energy sources and could result in a decrease in lignocellulose degradation. This bacterially induced retardation of fungal lignocellulose decomposition has, in fact, been shown to occur in an experiment on degradation of wheat straw by the white rot fungus Dichomitus squalens[72]. The same study, though, showed that the presence of bacteria did not hamper lignocellulose degradation by a Pleurotus sp. This was ascribed to the production of bacterial inhibiting compounds by the fungus, which is in agreement with the results of other studies reporting bactericidal effects of Pleurotus sp. in soils [73, 74]. Indeed, in situ bactericidal effects are likely to have evolved widely in the fungal kingdom as discussed earlier (Section 2.1). In addition to bactericidal effects, fungi have other mechanisms to prevent bacterial exploitation of released oligomers. For example, the highly hydrophobic nature of the mycelium of Pleurotus sp. appeared to prevent bacteria from penetrating into the straw substrate [72]. Another strategy may be the extra production of hydroxyl radicals, as has been shown for the brown rot fungus Antrodia vaillantii[75]. Additionally, acidification of the environment by fungal exudation of strong organic acids, e.g. oxalic acid, will prevent many bacterial strains from becoming active [51].

Although the aforementioned studies clearly show that competitive interactions between fungi and bacteria can be important during fungal decomposition of recalcitrant organic matter, the interactions are not necessarily always competitive. If fungal growth is constrained by factors other than the availability of soluble carbohydrates, bacterial consumption of these compounds may not be negative. For example, in studies on the effect of bacterial co-inoculation into spruce wood blocks with white rot fungi (Heterobasidion annosum, Resinicium bicolor, or Hypholoma fasciculare), the degradation of the wood consistently tended to be higher in co-inoculated trials than in trials inoculated with the fungus alone, even though wood decay could not be ascribed to the bacteria by themselves [76]. This stimulation may simply have been due to bacterial production of growth factors needed by the fungi, especially vitamins. It may also have related, however, to increased fungal enzyme activity stimulated by bacterial utilization of a proportion of the breakdown products. Relatively high levels of small-molecule breakdown products would tend to cause down-regulation of fungal enzymatic activity, but bacterial activity may reduce concentrations sufficiently to prevent this moderating effect.

Otherwise, bacteria may positively affect fungal activity by producing cellulase and pectinase enzymes that increase accessibility of substrates to the fungus. In addition, bacteria may also decompose solutes that are toxic to particular fungi. Finally, the activities of organotrophic, nitrogen-fixing bacteria may increase the nitrogen levels available for fungal growth [77–79].

3 Bacterial niches related to the utilization of fungal-derived substrates

The development of fungi in terrestrial ecosystems has resulted in a loss of some potential niches for bacteria, but it has also created opportunities to establish new niches. The basis of these niches is bacterial consumption of substrates derived from fungi.

3.1 Fungal-exudate consuming bacteria

Plate counts and in situ microscopic observations have both revealed the presence of bacteria on the surfaces of fungal hyphae and spores, as well as on mycorrhizal roots and in fruit body insides [80–88]. It is widely assumed that fungal exudates are a major or exclusive source of nutrients for these bacteria [84, 89–94]. This assumption underpins the idea that distinct habitats exist, which are characterised by enhanced or altered microbial activity in the soil around fungal structures. For example, the soil around mycorrhizas is considered to be sufficiently under fungal influence that a mycorrhizosphere microhabitat can be distinguished from the rhizosphere microhabitat found around non-mycorrhizal roots [95, 96]. Likewise, soil immediately adjacent to extraradical mycorrhizal hyphae and non-mycorrhizal hyphae is considered to subtend a mycosphere habitat distinct from the bulk soil.

Most studies on the composition of bacterial communities on fungal surfaces have been done for agricultural and economically important fungi, such as mycorrhiza formers, pathogens and saprotrophs that produce edible fruit bodies. Such studies have been largely restricted to the culturable fraction of the bacterial community. Strains closely related to known species of the genera Pseudomonas, Burkholderia and Bacillus are dominant among these culturable inhabitants of fungal surfaces (Table 2); former works in which slightly outdated taxonomies are utilized are also consistent with this picture [81]. In addition, novel fungus-associated culturable species have also been detected [97, 98]. The use of molecular techniques has detected both culturable bacteria, e.g. bacilli, and non-culturable Archaea [99, 100]. Although culture techniques may have overestimated the importance of pseudomonads and bacilli [87], it is obvious that these bacteria are frequent inhabitants of fungal surfaces. Not surprisingly, therefore, to date almost all studies aiming to elucidate the relationship between fungi and associated bacteria have been restricted to these bacterial groups.

2

Examples of bacteria associated with fungal surfaces

Fungal species	Surface	Isolation/counting method	Identification method	Dominant bacteria	Extra information	References
Suillus grevillei (EM)	Sporocarps (interior)	Plating	Physiological test	Pseudomonas, Bacillus and Streptomyces		[195]
Lactarius rufus (EM)	Mycorrhizal root (surface sterilised)	Plating	Biochemical test, FAME, 16S rDNA	Burkholderia, Pseudomonas and Paenibacillus	EM on Pinus sylvestris	[129]
Glomus clarum (AM)	Spores	Plating	FAME	Bacillus, Pseudomonas and Burkholderia	Only Bacillus after surface decontamination	[115]
Unidentified AM-fungi and Glomus dussii	Hyphae	Immuno-capture and plating	16S rDNA	Bacillus, Paenibacillus and Arthrobacter	gfp-tagging of Bacillus cereus	[99]
Suillus luteus (EM)	Mycorrhizal root	Plating	Physiological test, 16S rDNA	Bacillus, Burkholderia and Pseudomonas	EM on Pinus sylvestris	[116]
Pleurotus ostreatus	Mycelium	Plating	16S rDNA	Many different genera	Fungi grown in cotton compost	[135]
Cantharellus cibarius (EM)	Sporocarps (interior)	Plating	Physiological test	Pseudomonas (fluorescent)	Similar results for Hydnum rufescens	[90]
Lactarius spp.	Mantle	Microscopy	FISH	α-, β-, γ-proteobacteria	EM on Fagus sylvatica	[87]
Tuber borchii	Sporocarps (interior)	Plating	Physiological test, 16S rDNA	Pseudomonas, Bacillus and Paenibacillus	Cellulolytic and chitinolytic bacteria	[196, 197]
Phanerochaete chrysosporium	Hyphae	Plating	FAME	Agrobacterium and Burkholderia	Fungal strains from culture collections	[110]

Abbreviations: FISH, fluorescence in situ hybridisation; FAME, fatty acid methyl esters.

3.2 Selection of fungal-exudate consuming bacteria

An important question to be addressed in the study of common culturable bacteria is whether there is a specific fungal selection for particular bacterial strains, a phenomenon that could indicate an established and ongoing fungal-bacterial association. Exudation of soluble fungal storage sugars (usually trehalose) and polyols (e.g. mannitol) has been suggested as a possible mechanism for selection of fungus-associated bacterial strains by ectomycorrhizal (EM) fungi [90, 92, 101]. Fluorescent pseudomonads associated with Douglas fir –Laccaria bicolor mycorrhizas were revealed by genetic fingerprinting to be different from those in the bulk soil [101]. The EM-associated strains were able to degrade trehalose (produced by the EM fungus), whereas this ability was rare among the bulk soil strains. Trehalose-degrading fluorescent pseudomonads were also dominant among culturable bacteria in fruit bodies of Cantharellus cibarius[90], but in this case, trehalase activity was also common for Pseudomonas strains in the bulk soil [102].

Organic acids may also be selective for fungus-associated bacterial strains. External hyphae of various EM fungal species release organic acids, in particular oxalic acid, which are thought to be involved in nutrient withdrawal from solid mineral substrates [103, 104]. Oxalic acid is also released by white-rot fungi during degradation of lignocellulose [103]. Although the capacity to degrade the highly oxidized oxalic acid or its salt, oxalate, is not common for soil bacteria, oxalotrophy is taxonomically widespread [105]. The genus Methylobacterium is frequently found as the dominant oxalotrophic bacterium on and near oxalate-exuding plants, while the genera Alcaligenes, Pseudomonas, Ralstonia and Streptomyces have been mentioned as important oxalotrophic bacteria in soils [105]. Streptomyces has been found in association with an oxalate-producing Douglas fir mycorrhiza [106]. No additional information appears to be available on associations between oxalotrophic bacteria and fungi, an area worthy to further examine in the future.

There is still little information available on the quality and quantity of carbon compounds exuded into the environment by fungi [92, 96, 107]. More evidence is clearly required to support the idea that there is a substrate-mediated increase and selection of bacteria by fungi, especially in light of a recent study finding that, when thymidine incorporation was used to quantify in situ bacterial activity [108], there was no support for the hypothesis that EM mycelia can stimulate bacterial growth via carbon exudation. Furthermore, for arbuscular mycorrhizal (AM) fungi, it has been suggested that the effect of exudates on bacterial populations may be qualitative (relating to species and strain composition) rather than quantitative [94].

Exudation of inhibitory chemicals by AM and EM fungi has been suggested as a mechanism for selection of fungus-specific, antibiotic-resistant bacteria [89, 108]. Some saprotrophic basidiomycetes are also known to produce antibiotics, i.e., various agaricoid species produced water-extractable antibiotics that inhibited Gram-positive bacteria but not the Pseudomonas spp. that had been isolated from the fruit bodies [109]. Since antibiotic production is likely to be induced in the presence of appropriate bacteria, the range of basidiomycetes found to produce antibiotics is likely to be greater when screening procedures include bacteria in the fungal cultures than it would be in a study of pure fungal cultures. Of course, secretion of antibiotics by fungi in pure culture does not necessarily mean that these antibiotics are also secreted in natural environments. In various studies, interesting fungus-associated bacterial strains have been detected precisely because of their resistance to broad-spectrum antibacterial antibiotics employed in isolation media for fungi. This demands that antibiotic-mediated selection of bacteria by fungi deserves more attention [98, 110].

As well as these direct effects of fungi on bacteria, indirect effects may also be seen. It has been suggested that mycorrhizal fungi indirectly influence bacterial communities in the mycorrhizosphere by stimulating root growth, as well as by altering root exudation patterns and the structure of the surrounding soil [93, 96, 111]. Such effects apply not only to free-living bacteria but also to those involved in root nodule symbioses. AM fungi have been shown to enhance nodulation and nitrogen fixation in legumes. The increased P uptake that occurs when these fungi are present is beneficial to the functioning of the nitrogenase enzyme of the bacterial symbiont [96, 112].

Although the aforementioned studies certainly do indicate the potential for specific fungal selection of bacterial strains, non-specific adherence of bacteria to fungal hyphae and spores has also been observed [113–115]. In addition, soil conditions have been shown to affect the composition of bacterial populations associated with fungi [116]. It would appear that the probability of a particular bacterial species being selected by a fungus is dependent on its abundance in the bulk soil bacterial community, with the same fungal species having different bacterial associates in edaphically different sites. Rhizosphere bacterial community composition has also been shown to depend on the bacterial community composition of adjacent bulk soil [117].

Besides offering the possibility of growth on fungal exudates, hyphae may be important for the transport of bacteria, particularly for those associated with plants. The nitrogen-fixing, endophytic bacteria of the genus Azoarcus that occur in grasses have been found in association with rhizosphere-inhabiting fungi [118, 119], suggesting that the bacteria colonize new plants via transport on or in fungal hyphae. Furthermore, the possibility of adherence of Rhizobium leguminosarum to hyphae of the AM-fungus Gigaspora margarita has been demonstrated [113]. The possible vector function of fungal hyphae does, however, imply active growth or movement of bacteria along the hyphae, as mature parts of hyphae do not move towards the roots but reach them via apical growth.

3.3 Effects of fungal-exudate consuming bacteria on fungal performance

Associated bacteria may have negative, neutral or positive effects on fungal fitness. There are several indications of mutualistic relationship between fungi and their associated bacteria, the most obvious being found in the cyanolichens. The fungi benefit from the relationship by obtaining a supply of energy from the cyanobacteria. Unlike the green algal photobionts in most lichens, some cyanobacterial partners fix nitrogen and supply a portion of the yield to the fungus [120–123].

Nitrogen transfer may also be important in the association of organotrophic bacteria with fungi. Nitrogen-fixing bacilli were associated with Douglas fir –Rhizopogon vinicolor“tuberculate” EM [124]. Respiration in the tuberculate mycorrhizal complex by the fungus, roots and associated microorganisms, probably contributed to maintaining a microaerophilic environment where nitrogen fixation could take place. The authors of this study hypothesized a mutualistic relationship in which the bacteria supplied to fungus with nitrogen while growing on fungal exudates. Nitrogen-fixing bacteria have also been found on hyphae in mycorrhizal structures of Arbutus unedo[125]. To date, however, no data have been reported that would shed light on the relative importance of nitrogen fixation by exudate-consuming bacteria for the nitrogen supply of mycorrhizal fungi.

The most well-known example of beneficial effects conferred by fungus-associated bacteria is the so-called EM helper bacteria [126]. The earliest studies concerned the positive effect on EM establishment by fluorescent pseudomonads that had been isolated from surface-sterilized EM of the same fungi, e.g. Rhizopogon luteolus and Laccaria spp. (reviewed by Garbaye [126]). Several follow-up studies have confirmed the stimulating effect of these bacteria on mycorrhizal establishment, but the actual mechanism has not been elucidated yet [127, 128]. Other studies have made it clear that the ‘helper’ phenomenon is not restricted to fluorescent pseudomonads as EM-associated bacilli and paenibacilli have also been shown to stimulate mycorrhizal formation [116, 129]. In addition, an EM-associated streptomycete was shown to stimulate growth of different EM fungi [130].

Bacterial stimulation of AM growth and plant infection has also been reported [91, 131, 132]. However, in most cases these reports concern so-called plant-growth promoting rhizosphere bacteria that have no clear association with AM fungi. Nevertheless, such bacteria are also sometimes referred to as helper bacteria [133].

Fungus-associated bacteria can also have effects on fruit body formation and spore production. Fruit body initiation of Agaricus bisporus is dependent on the presence of Pseudomonas putida bacteria. The exact mechanism has yet to be determined but it has been suggested that the mycelium produces self-inhibiting compounds, which have to be removed by the bacteria [134]. P. putida strains were also shown to promote mycelial growth and fruit body formation of Pleurotus ostreatus[135]. However, care should be taken when interpreting the outcome of these interactions as beneficial to the fungus, since fruiting and increased hyphal extension rate could equally well be interpreted as a stress response.

Several cases of stimulation of spore germination by spore-associated bacteria have been reported [115, 136–138]. The actual mechanism(s) is not known but production of germination-inducing volatiles, degradation of germination-inhibiting compounds, and enzymatic weakening of the spore wall have been proposed. The last-mentioned possibility is supported by electron micrographs of spore-wall-eroding activities of chitinolytic bacteria [139, 140].

In contrast, inhibition of spore germination by spore-associated bacteria has also been frequently reported [115, 141, 142]. In fact, the restriction of germination of most fungal spores in non-amended bulk soil is a well-known phenomenon, referred to as fungistasis (or mycostasis), which has been attributed to withdrawal of nutrients from the fungal spores by associated bacteria [141, 143] and to the production of bacterial antibiotics [144, 145].

3.4 Endosymbiosis

The occurrence of bacteria inside tissues or cells of eukaryotes has been reported for plants, animals and protists [146–148]. The status of these bacteria varies from occasional co-existence to obligate dependency on the host. Indeed, mitochondria and chloroplasts may be regarded as the ultimate obligately dependent bacteria that have lost most of their original bacterial properties [149, 150].

To date, comprehensive studies on the occurrence of endocellular bacteria in fungi are lacking. A notable exception is the research of Bonfante and her colleagues on such bacteria in AM fungi [151, 152]. Earlier microscopic studies by different research groups had already revealed the presence of bacteria-like organisms (BLOs) inside spores and hyphae of AM fungi (reviewed by Scannerini and Bonfante-Falso [152]). However, as these BLOs appeared to be unculturable, it was not until the advent of PCR analysis that their bacterial status could be confirmed and their phylogenetic position determined [153, 154]. Most detailed studies have been done on the endocellular bacteria of Gigaspora margarita. Sequence analysis of the 16S rDNA revealed that the bacteria formed a distinct lineage within the β-proteobacteria, with the genera Burkholderia, Pandoraea and Ralstonia as closest neighbours [154]. Phylogenetically closely related bacteria were found in G. margarita isolates from different geographical regions as well as in two Scutellospora species [155]. In contrast, endocellular bacteria were not detected in different isolates of Gigaspora rosea.

The nature of the relationship between endocellular bacteria and AM fungi is not clear. There were no indications that the bacteria had any negative effect on the symbiotic efficiency of the AM fungus [156]. One study suggested that the expression of nif (nitrogen fixation) genes in germinating AM spores could indicate that the bacteria are supplying the fungus with nitrogen during its pre-infective growth [157]. However, subsequent studies have shown that this is probably not the case (Bonfante, personal communication).

The close phylogenetic relationship of the endocellular bacteria in geographically separated AM strains and in different AM species may indicate that bacterial acquisition by these fungi was a unique event that occurred in an ancestral fungus and has since then been propagated by vertical transmission [155, 158]. However, more data on the identity of BLOs of other AM species are needed before this idea can be fully evaluated.

Obviously, uptake of bacteria by fungi is a less common event than is uptake by protozoan bacterial predators, which utilize phagocytosis as a mechanism for incorporating the bacterial cells. The fungal cell wall as a physical barrier strongly tends to prevent acquisition of bacteria [155]. The wall is not rigid at the hyphal tip [159], and bacterial acquisition may occasionally occur at this location. Such an event does, in fact, occur with the fungus Geosiphon pyriforme, which is probably a representative of a lineage ancestral to AM fungi [160]. This fungus incorporates primordia of free-living cyanobacteria (Nostoc) at the hyphal tips, which thereafter swell and form bladders in which the Nostoc cells initiate photosynthetic activity [161].

Another scenario for bacterial acquisition by fungi could be the lysis of hyphal tips by bacteria, followed by entry into the fungal mycelium (see Section 3.5). Related to this, a recent work reporting on invasion of AM fungal spores by Burkholderia sp. is very interesting [162]. These bacteria appear to invade germinating spores via germ tubes that were weakened either by the absence of a host plant or by the action of bacterial lytic enzymes. In nature, similar events may occur; since hyphae from germinating AM spores have been shown to undergo programmed growth arrest and resource reallocation as a strategy to maintain germination capacity in the absence of a plant host [163]. The insides of such retracting hyphae may be readily accessible to bacteria [162]. It is possible that this type of event has occasionally been the first step in the evolution of a lineage of endocellular bacteria.

If ‘fungal capture’ of lytic bacteria, or bacteria that enter damaged hyphae, is not uncommon, then the possession of endocellular bacteria by fungi may be a widespread phenomenon. The increasing number of reports on the occurrence of endobacteria in fungi, other than AM, seems to indicate that this may actually be the case [84, 164–167]. Interestingly, the only bacteria that have been indicated as endobacteria of EM fungi belong to genera with the potential to hydrolyse fungal cell wall polymers, e.g. Paenibacillus and Cytophaga [164, 166].

3.5 Mycophagy

Besides being competitors or suppliers of exudates, living fungi may also serve directly as sources of nutrients for other microorganisms. A nutritional strategy based on targeting fungi as a substrate is well known for the so-called mycoparasitic fungi. Several studies have addressed the mechanism and regulation of attack on fungi by Trichoderma spp. [168–170]. The following steps can be recognized: (1) chemotropic growth towards the host fungus, (2) coiling around the host fungus and appressorium formation, (3) secretion of cell wall degrading enzymes, (4) penetration and (5) degradation of hyphal content [168]. In a previous study, coiling around and penetration of a host fungus by a Streptomyces strain was also described [171]. However, with the exception of studies on penetration and killing of living fungal spores no further attention appears to have been paid to the occurrence of mycophagy in actinomycetes [172].

Lysis of fungal hyphae by non-filamentous bacteria has, however, been observed frequently [173–177]. Studies on this topic concern the deformation or destruction of hyphae of plant pathogenic soil fungi by potential biocontrol bacteria. However, due to the artificial conditions used in most experiments, e.g. use of liquid nutrient broth and/or the use of damaged mycelium as inoculum, it is often not clear whether the bacteria are just feeding on nutrient broth and dead or damaged fungal material or are really able to attack intact fungal hyphae. Even if such an attack is found, it is not evident if it would also occur in a nutrient-limited soil or soil-like environment. These studies have revealed how fungal attack by unicellular bacteria might proceed. The fact that almost all mycolytic bacteria produce polymer hydrolysing enzymes (e.g., chitinases, glucanases or proteases) as well as antibiotics, indicates that this mixture of compounds may be crucial for degrading living hyphae. Microscopic observations indicated that the initial bacterial attack proceeded via polar attachment of bacteria to hyphae, followed by biofilm formation on the hyphal surface [177, 178].

Recently, a new bacterial genus, Collimonas, has been described to be able to grow at the expense of living hyphae of different fungi in soil microcosms [179, 180]. These bacteria can be dominant among the culturable, chitinolytic bacteria in fungal-rich, acidic or sandy soils. The Collimonas strains appear to attack the hyphal tip using a combination of lytic enzymes (chitinases) and antibiotics. However, the exact mechanism of this attack, as well as the relative importance of mycophagous growth, has yet to be elucidated for these bacteria.

Myxobacteria are also likely to be mycophagous in soil. These Gram-negative bacteria have been mostly studied because of their social behavior during collective food uptake, as well as the cooperative motility involved in their fruit body formation [181]. They have been termed micropredators due to their ability to destroy living soil microorganisms, in particular bacteria and yeasts [182]. They produce various exoenzymes and secondary metabolites that appear to be involved in these processes. Up till now, their ability to grow on filamentous soil fungi has not been studied intensively. In vitro studies have revealed hat they can antagonize plant pathogenic fungi as well as saprotrophic fungi [183]. Myxobacteria were indicated as responsible for perforation of Rhizoctonia sp. hyphae that had been buried in soil [184]. In this study, myxobacteria isolated from the buried hyphae were also shown to penetrate and to empty out Rhizoctonia sp. hyphae in vitro.

Another group of bacteria for which mycophagy may be important is the paenibacilli. These bacteria are known for their extracellular lytic enzyme production and have been shown to attack fungi in vitro [177, 185]. In addition, they are often isolated from the surfaces or even insides of fungal propagules [88, 116, 162, 164].

Bacterial pathogens of fungi may be considered as a special group of mycophagous bacteria. The best-studied case is that of brown blotch disease of the cultivated mushroom (Agaricus bisporus) by Pseudomonas tolaasii[186]. P. tolaasii grows as an organotrophic bacterium in soils and composts but under certain conditions it can attack hyphae in fruit bodies, where the zones of attack can be recognized as brownish spots. Several extracellular secondary metabolites, of which tolaasin appears to be the most important, are involved in membrane disruption releasing nutrients from within the fungus cells [186]. This attack initiates a cascade of events culminating in the production of chemical barriers (melanins, quinones) by means of which the fungus attempts to protect the cells in the inner parts of the fruit body from becoming infected.

Not only are there bacteria that feed on fungi, but also there are fungi that are able to lyse and consume bacteria [187]. It has been suggested that bacteria may be an important source of nitrogen during fungal degradation of resources with a high C/N ratio, e.g. wood [188]. Fungi with the ability to lyse bacteria appear to be attracted by bacterial colonies. It has been speculated that density-dependent control of antifungal gene expression, which is apparent for several soil bacteria, could be a strategy to defend bacterial colonies against fungal attack [189]. This idea, however, has not been experimentally examined yet.

3.6 Decomposition of fungal cell walls

Actinomycetes have long been considered to be the major degraders of senescent fungal hyphae [190]. This idea was primarily based on microscopic observations of the colonization and degradation of fungal hyphae growing on surfaces. It is known, though, that the use of surfaces, such as glass slides, may introduce artefacts that influence these events. On the other hand, it has been shown frequently that stimulation of fungal growth by substrate addition is later followed by an increase in actinomycete densities after the decline of the fungal activity [191]. In addition, the growth of potworms (Enchytraeus albidus) feeding on fungi was enhanced by gut-inhabiting streptomycetes that degraded the fungal cell walls [192]. Hence, it is certainly possible that actinomycetes play an important role as degraders of senescent fungal mycelium, but more research on the relative significance of different organisms is needed. Besides actinomycetes, the soil dwelling myxomycetes may also be important degraders of senescent fungal mycelium [182]. Moreover, in wood, when one fungus replaces another during community development [15], it is highly likely that both intracellular and wall materials are used by the invading organism, which is often effectively a pure culture, in order to satisfy nitrogen demands.

4 Perspectives

It will be clear to every soil microbiologist that interactions between plants and bacteria have had a strong impact on evolution and niche differentiation of terrestrial bacteria. It is much less widely appreciated that terrestrial bacterial niche development may for a large part have been determined by the development of fungi on land. Consideration of this scenario may give rise to new ideas and concepts in soil microbial ecology.

The increasing availability of both fungal and bacterial genome sequences will undoubtedly provide a better picture of the impact that fungi have had on bacterial evolution. Of particular interest in this respect are horizontal gene transfer between fungi and bacteria (and vice versa), evolution of antifungal compounds, detailed comparison of fungal and bacterial genes involved in cellulose degradation, and comparison of genomes of fungus-associated bacterial strains with those of phylogenetically related free-living strains.

Investigation of fungally engendered bacterial niches has only recently started, and has already provided exciting findings such as mycorrhiza-helper bacteria, endobacteria and mycophagous bacteria. Other areas, such as interactions during degradation of recalcitrant organic matter, are almost entirely unexplored. With respect to the last topic, a further extension of such investigations to aquatic environments is necessary as there is increasing evidence that interactions between bacteria and fungi occur during degradation of terrestrially derived organic matter, e.g. leaves, that are sedimented into lakes and seas [193, 194].

The study of fungal-bacterial interactions in soils is not only interesting from a basic point of view but has also yielded findings of societal and economical relevance, such as the application of bacteria for the biocontrol of fungal plant diseases, the improvement of mushroom cultivation procedures, and the controlled stimulation of mycorrhizal infection. Better insight into the co-existence mechanisms of soil bacteria and fungi may lie at the basis of both new and improved applications. Bacteria that are harmful to fungi may potentially also form a source of new antibiotics with therapeutic value. A better basic understanding of the in situ competitive interactions between bacteria and fungi will probably result in much more straightforward approaches than currently exist to screen for new antibiotics and other valuable secondary metabolites.

In summary, there is much to be gained from studying fungal–bacterial interactions and it is encouraging to see that an increasing number of research groups have started to explore this fascinating area of microbial ecology.

⋆

This is publication 3466 of NIOO-KNAW Netherlands Institute of Ecology.