Wars between microbes on roots and fruits

Niche colonization

Root colonization: competition for nutrients and niches. A competitive root tip colonization assay was developed by Simons et al.¹². Microbiologically sterile seeds are coated with a mixture of cells of two different strains, and then each coated seed is placed in a sterile system with a plant nutrient solution and allowed to germinate. When the root is approximately 10 cm in length, the number of microbes at various locations along the root is counted. These numbers are highest at the root base and decline rapidly towards the root tip. The ratio of the two types of microbes at the root tip appears to be a sensitive way to determine which strain is the best competitive colonizer¹².

The ability for bacteria to occupy specific sites on the root is a major route to ensure close proximity to the high concentrations of secreted nutrients. Many genes and traits required for efficient competitive root colonization have been identified (reviewed in 13–15). The best-understood ones are the following: (i) chemotaxis towards root exudate compounds, where L-isoleucine was identified as the strongest chemoattractant in tomato root exudate but, when corrected for the quantities present, malic acid and citric acid are most effective¹⁶; (ii) fast growth on root exudate components¹⁷; and (iii) the type three secretion system, whose function involves pinching a needle through the plant cell membrane in order to extract nutrients from that source¹⁴.

Chin-A-Woeng et al.¹⁸ have shown that root colonization is essential for the control of the disease tomato foot and root rot (TFRR) caused by the fungal pathogen Fusarium oxysporum f. sp. radicis-lycopersici (Forl). This conclusion was based on the inability of three different competitive colonization mutants to efficiently colonize the root tip using the assay described above. Defective colonization was confirmed by experiments in soil.

The method of Simons et al.¹² was used by Kamilova et al. as a tool to enrich extremely efficient competitive root colonizers. They inoculated seeds with a mixture of total tomato rhizosphere bacteria. After root growth, they cut off the root tip, containing far less than 1% of the total number of bacteria present on the root. This root tip sample was intended to be enriched in bacterial strains that are superior competitors for exudate nutrients and for rapidly reaching the root tip. After two more enrichment cycles, whereby the root tip population was serially reinoculated via sterile seeds, individual root tip bacteria were purified and investigated. Most isolates were pseudomonads, and approximately half were superior in competition experiments for reaching the root tip niche after application on a sterile seed. In competition growth experiments in sterile exudate, the enhanced colonizers were at least 20-fold more effective than the best bacteria randomly selected from the starting mixture. Most importantly, four of the five tested enhanced colonizers appeared to control TFRR. So, it is clear that this method enriches for bacteria that exert biocontrol on roots via competitive exclusion. So far, this is the only class of biocontrol bacteria for which an enrichment method exists. This biocontrol mechanism was designated as competition for nutrient and niches (CNN)¹⁹.

Another way to rapidly generate enhanced colonizers is to apply cells of a mutY mutator strain on sterile roots, followed by the enrichment procedure mentioned above. By this approach, rather than enriching for natural isolates, bacteria reaching the root tip have fixed a set of mutations that evolutionarily enhance their competitive colonization ability. The sets of mutations required for this increased colonization differ for the dicot tomato and the monocot grass²⁰; however, as yet, the individual effects of the fixed mutations remain uncharacterized.

Colonization of fungal hyphae during biocontrol. Colonization of Forl hyphae has been illustrated for biocontrol strains Pseudomonas fluorescens WCS365 and Pseudomonas chlororaphis PCL1391 by confocal laser scanning microscopy²¹. De Weert et al.²² discovered that the bacterial cells move toward the hyphae by using fusaric acid (FA) secreted by the hyphae as the major chemoattractant.

Kamilova et al.²³ observed that WCS365 cells colonize developing Forl hyphae and that enhanced colonization of the hyphae results in the formation of microcolonies (Figure 1) and correlates with low nutrient availability. These results suggest that hyphal colonization is used to move to and occupy a high resource niche, e.g. fungal exudate. Colonization by the bacteria does not kill or deform the hyphae but results in slower hyphal growth. For further details on the effects of P. fluorescens WCS365 on Forl hyphae and spores, see also the section titled “Interference with activity, survival, germination, and sporulation of the pathogen”²³.

Figure 1. Scanning electron microscopy of colonization of Fusarium oxysporum f. sp. radicis-lycopersici (Forl) hyphae by cells of Pseudomonas fluorescens WCS365 after incubation for 72 hours in tomato root exudate.

The right photo is a detail from the left photo but with a higher magnification (the arrow indicates a bacterial microcolony). Reproduced from Kamilova et al.²³ by permission of the publisher.

Hyphal colonization by other parasitic fungi has also been observed. Upon contact with fungi, the biocontrol fungus Trichoderma harzianum attaches to the host plant pathogenic fungus and can coil around it. This so-called hyphal mycoparasitism is suggested to be the mechanism by which T. harzianum controls Sclerotinia sclerotiorum in lettuce and sunflower seedlings²⁴.

Molecular Weapons

The microbial arsenal is exceptionally diverse, with different weapons deployed in response to dangers of different magnitude and proximity. Thus, while the effects of secreted diffusible products like antibiotics, enzymes, siderophores, and bacteriocins are highly localized to the producing organism, the influence of volatile weapons is likely to be significantly broader²⁵. We briefly examine these differences below.

Antibiotics. Examples of antibiotics used by Gram-negative biocontrol bacteria in the rhizosphere to attack fungal plant pathogens include 2,3-deepoxy-2,3-didehydro-rhizoxin²⁶, phenazine-1-carboxylic acid (PCA), phenazine-1-carboxamide (PCN), 2,4-diacetylphloroglucinol (Phl), pyrrolnitrin, 2-hexyl-5-propyl resorcinol, and the cyclic lipopeptide viscosinamide^1,6,8,9 (for structures, see 27). It should be noted that strains FZB42²⁸ and SQR9²⁹ of the Gram-positive biocontrol bacterium Bacillus amyloliquefaciens are also sources of many antibiotics, such as fengycins, iturins, and surfactins^28,29. Also, many fungi produce antibiotics, which can be active against bacterial and/or fungal plant root pathogens^15,30,31.

Siderophores. Iron is an abundant element on the earth crust, but it is insoluble and therefore not suitable for uptake by living organisms. The free Fe³⁺ concentration is only approximately 10^-18 M. Microbes, like all organisms, need Fe ions. In order to scavenge Fe³⁺ ions, they secrete siderophores, which are Fe³⁺ ion-chelating molecules. Upon binding the ion, the formed siderophore–Fe³⁺ complex is bound by iron-limitation-dependent receptors at the bacterial cell surface, and the Fe³⁺ ion is subsequently released and becomes active in the cytoplasm as Fe^2+1,15. Examples of siderophores are pyocyanin and pyoluteorin (for structures, see 27).

The superior ability of certain bacteria to scavenge Fe³⁺ ions can result in a reduction of the growth of pathogens. Examples of this form of biocontrol are the control of Erwinia carotovora (renamed as Pectobacterium carotovorum) by P. fluorescens strains³² and the control of pathogenic fungi^33,34. This form of biocontrol is most efficient when the Fe³⁺ concentration is low, e.g. in acidic soils³⁴. An intriguing phenomenon is that similarities in siderophore structure can allow some species of bacteria to “pirate” those of others (called xenosiderophores), leading to iron deficits and growth or developmental arrest^35–37. Since such xenopiracy also occurs within the rhizosphere, this offers interesting possibilities for biocontrol^33,34,38.

Bacteriocins. Bacteriocins are structurally and functionally diverse antimicrobials that are generally distinguished from antibiotics by their narrow spectrum of activity against strains of the same species or congeners. Because of this narrow target range, bacteriocins are assumed to exert their strongest influence against direct competitors within the same ecological niche as the producing strain, a feature that makes them especially attractive for the purposes of biocontrol. Given the apparent ubiquity of bacteriocins – essentially all bacterial species carry a diverse set – we will only touch on this topic here, referring instead to several excellent reviews^39–41.

Within the rhizosphere, pseudomonads produce strain-specific pyocins that vary in size and cellular target, designated R, S, and F-type bacteriocins. While S-type pyocins are related to structurally similar small peptide colicins from Escherichia coli, R- and F-type bacteriocins are evolutionarily derived from defective phages⁴². In addition, pseudomonads including the well-known biocontrol strain P. fluorescens Pf-5 (renamed Pseudomonas protegens) produce lectin-like bacteriocins called putidacins or L-type bacteriocins, with genus-specific antimicrobial activity owing to the targeting of D-rhamnose within their cell walls⁴³. Importantly, putidacins have shown efficacy in field trials against Pseudomonas syringae, the causative agent of olive knot disease⁴⁴. Within soil and the rhizosphere, Bacillus bacteria also produce a vast array of bacteriocins with potential application to biocontrol. Notably, Bacillus bacteriocins seem to have broader target ranges, even extending to Gram-negative species⁴⁵. For example, a bacteriocin from Bacillus subtilis 14B is active against Agrobacteria, while another strain of the same species, IH7, reduces the incidence of damping off and seed-borne diseases when applied as a seed coating^46,47.

Despite the promise of bacteriocins for biocontrol, it is important to bear in mind that the microbial targets of these agents are neither evolutionarily static nor defenseless. As with antibiotics, bacteria can readily evolve resistance to bacteriocins, thereby limiting their long-term efficacy. Equally, pathogenic bacteria produce many of the same bacteriocins as the biocontrol strains that are used to kill them. Indeed, bacteriocins are predicted to be one of the central factors driving bacterial diversification, an idea borne out by the tremendous diversity of bacteriocin loci within species^48–50.

Enzymes. Several soil-borne bacteria and fungi produce extracellular enzymes such as cellulases, chitinases, β-1-3 glucanases, lipases, and proteases. These lytic enzymes can hydrolyze a wide variety of polymeric fungal cell wall compounds, including cellulose, chitin, hemicellulose, and protein, thereby interfering with pathogen growth and/or activities. Production and secretion of these enzymes by different microbes can result in biocontrol. This mechanism, predation and parasitism, is a mechanism used by some biocontrol species of the fungus Trichoderma⁵¹. Predation and parasitism is less well-known as a biocontrol mechanism used by bacteria, although a few examples exist⁵². It should be noted that predation and parasitism is not always related to biocontrol. For example, the biocontrol strain Collimonas fungivorans is famous for eating live fungi, but its major mechanism of action for controlling TFRR is CNN⁵³.

Interference with activity, survival, germination, and sporulation of the pathogen. Kamilova et al.²³ studied the interaction of WCS365 cells with Forl (see section titled “Colonization of fungal hyphae during biocontrol” and Figure 1) also in the presence of tomato root exudate. They observed that tomato root exudate stimulates fungal spore germination, an effect that can be mimicked by plant growth solution supplemented with citrate or glucose²³, the major organic acid and sugar compounds of exudate⁴. However, the presence of WCS365 cells in exudate reduces spore germination and subsequent development of hyphae. Although the precise mechanisms remain uncertain, the authors suggest that the inhibition of spore germination by WCS365 is the result of consumption of nutrients that act as germination inducers²³. Furthermore, after the growth medium was limited in nutrients, abundant hyphae colonization by WCS365 cells was observed, supposedly as a result of their search for nutrients exuded by the fungus. Simultaneously, a reduction in microconidia production was observed. It therefore can be concluded that the reduction of spore germination, the colonization of developing hyphae (Figure 1), and the subsequent inhibition of the production of new spores diminish fungal vigor as well as fungal dissemination, thereby contributing to the biocontrol of TFRR by WCS365 bacteria²³ whose major mechanism of biocontrol is induced systemic resistance (ISR) in the plant¹⁹. See the following section for explanation of the latter mechanism.

Indirect competition via induction of resistance in the plant. Interactions between selected plant growth-promoting bacteria and fungi with plant roots can prime these plants for resistance to a broad range of pathogens and insect herbivores while providing a competitive advantage for the inducing species. This biocontrol mechanism is called ISR and was recently reviewed by Pieterse et al.⁵⁴. The inducing microbes trigger a reaction in the plant roots that gives rise to a signal that spreads systemically throughout the plant and enhances the defensive capacity of distant tissues to subsequent infection by the pathogen. Examples of such microbes are species of the bacteria Bacillus and Pseudomonas and of the fungus Trichoderma and of mycorrhizal fungi. It should be noted that ISR differs from systemic acquired resistance, which is caused by a hypersensitive response triggered by plant pathogens.

In contrast to the situation with many other biocontrol mechanisms, extensive colonization of the root system is not required for ISR, as shown by experiments with P. fluorescens WCS365⁵⁵ using root colonization mutants. Not only whole cells but also many individual bacterial determinants can induce ISR, such as lipopolysaccharide, flagella, salicylic acid, the siderophores pyochelin and pyocyanin, some cyclic lipopeptides, the antifungal factor Phl, the signal molecule acyl homoserine lactone (AHL), and volatile blends as well as the individual volatiles acetoin and 2,3-butanediol^1,27,54.

Does malic acid of root exudates attract beneficial bacteria to the root? Rudrappa et al.⁵² found that Arabidopsis thaliana seedlings whose leaves were infected with the foliar pathogen P. syringae pv. tomato DC3000 demonstrated enhanced root secretion of L-malic acid; in addition, they suggest that this elevated L-malic acid level selectively signaled and recruited the beneficial biocontrol rhizobacterium B. subtilis FB17, which defends the plant via ISR. Previously, De Weert et al.¹⁶ reported that another biocontrol bacterium, P. fluorescens WCS365, which also defends the plant through ISR¹⁹, shows strong chemotaxis toward the major tomato root exudate components, including malic acid. Although the suggestion of Rudrappa et al.⁵² is intriguing, it seems unlikely that enhanced L-malic acid secretion can selectively attract beneficial bacteria given that chemotaxis to L-malic acid is found in both beneficial and potentially harmful bacteria.

Volatile organic compounds. VOCs, complex mixtures of low-molecular-weight compounds, are produced by many organisms. Maffei et al.⁵⁶ proposed the name “volatilomes” for VOC mixtures. More than 1,000 bacterial VOCs have been described so far, but the diversity of environmental niches suggests that this is a gross underestimation^57,58. Audrain et al.⁵⁷ divided bacterial VOCs into seven chemical classes. VOCs can travel far away from the site of production through the atmosphere, as well as through porous soils and liquids, making them ideal info-chemicals for mediating short- and long-distance interactions⁵⁶.

Microbial VOCs can exert a wide range of activities including controlling bacterial and fungal plant pathogens^57,59–61, signaling⁶², inhibiting microbial activity^57,63 and microbial growth^57,60,64, modifying drug resistance^57,63,65,66, e.g. by raising the pH of the culture medium⁵⁷, negatively affecting biofilm formation^57,67, eliciting ISR in a plant^68–73, eliciting induced systemic tolerance to stresses caused by drought and heavy metals^60,69, and promoting plant robustness⁷⁴ and plant growth^68,75.

Excellent reviews about the chemical diversity and structures^57,76 of VOC-producing microbes^60,76–78 and about the perspectives for application in sustainable agriculture and post-harvest control^60,79–82 have recently appeared.

VOC-mediated interactions between microorganisms can be easily demonstrated by using an I-plate, which is a Petri dish that contains two physically separated compartments that allows free exchange of air⁶⁶. However, the identification and quantification of VOCs requires special equipment^57,77,83. The volatile mixture can be analyzed using a combination of chemical profiles built by liquid chromatography–high resolution mass spectrometry or nuclear magnetic resonance and multivariate data analysis⁸³.

In the following paragraphs, we will discuss selected examples of activities of VOCs produced by microbes.

Volatilomes of many bacteria inhibit the mycelial growth of the plant pathogen Rhizoctonia solani⁷⁷. The strongest inhibition was found for volatiles of Stenotrophomonas maltophilia R3089, Serratia plymuthica HRO-C48, Stenotrophomonas rhizophila P69, Serratia odorifera 4Rx13, Pseudomonas trivialis 3Re2-7, S. plymuthica 3Re4-18, and B. subtilis B2g. Volatiles of strains of P. fluorescens and Burkholderia cepacia were moderately active. The VOC profiles of these antagonists differ in composition and complexity. Most volatiles are species specific, but overlapping volatile patterns were found for Serratia spp. and Pseudomonas spp.⁷⁷.

Using microarray analyses of E. coli exposed to the VOCs 2,3-butanedione and glyoxylic acid, Kim et al.⁶³ observed that these volatiles mediate global changes in gene expression related to motility and antibiotic resistance. They suggest that bacteria use airborne VOCs to sense other bacteria and to change master regulatory gene activity to adapt⁶³.

Raza et al.⁸⁴ reported that P. fluorescens WR-1, a biocontrol strain, produces Ralstonia solanacearum growth-inhibiting VOCs. Benzothiazole and 1-methyl naphthalene were the most effective at inhibiting growth, and the VOCs of P. fluorescens inhibited swarming, swimming, and chemotactic motility. These results suggest that the VOCs of the biocontrol strain, once established on the tomato root, not only reduce the growth of the pathogen but also inhibit movement of the pathogen towards the rhizosphere. Proteomics analysis showed 9 up-regulated and 19 down-regulated proteins. The latter include proteins involved in anti-oxidant activity, virulence, and carbohydrate and amino acid biosynthesis. The authors conclude that the results not only provide more insight into the mechanism of biocontrol but also could lead to the development of safe fumigants.

Many bacterial strains produce the antimicrobial VOCs ammonia and cyanide^57,58. Biogenic ammonia modifies antibiotic resistance at a distance in physically separated bacteria⁶⁵. The cyanide ion is a potent inhibitor of many metalloenzymes, especially copper-containing cytochrome c oxidases⁸⁵. In addition to a role in the control of plant diseases⁶, cyanide may play a role in increasing the availability of nutrients for plants through its ability to sequestrate metal ions⁸⁶. Dimethyl sulfide, produced by the rhizospheric bacteria P. fluorescens and S. plymuthica, has bacteriostatic effects on the plant pathogens Agrobacterium tumefaciens and Agrobacterium vitis ⁸⁷.

One of the best-studied volatiles is 2,3-butanediol⁷⁵. This volatile elicits plant growth promotion⁷⁵ and ISR⁸⁴ in A. thaliana, as was shown using mutants blocked in the synthesis of 2,3-butanediol and its precursor acetoin⁸⁸. More recently, Ryu’s group studied rhizosphere colonization of pepper, using the wild-type 2,3-butanediol producer B. subtilis 168, its null mutant, and a 2,3-butanediol-overproducing derivative. The results showed that a higher level of 2,3-butanediol production correlates with improved rhizosphere competence of B. subtilis itself and that 2,3-butanediol production by B. subtilis suppresses the growth of the saprophytic fungus Trichoderma in the rhizosphere⁷⁴. Furthermore, growth experiments were carried out in exudates of roots pre-treated with the 2,3-butanediol-overproducing strain. This pre-treatment negatively affected the growth of the B. subtilis null mutant. This indicates that 2,3-butanediol protects B. subtilis (directly or indirectly) from harmful exudate components. The same pre-treatment also inhibited the growth of the soil-borne bacterial pathogen R. solanacearum but enhanced the growth of the saprophytic biocontrol bacterium P. protegens Pf-5. Apparently, 2,3-butanediol alters the root exudate composition⁷⁴.

The volatile 2,3-butanediol triggers strong ISR against necrotrophic bacteria but not against biotrophic ones. Park et al.⁸⁹ observed that the C16 volatile hexadecane confers protection against both types of pathogens. The C13 volatile tridecane, emitted by Paenibacillus polymyxa, was shown to be a more powerful inducer of ISR than was 2,3-butanediol⁷⁰.

Defense strategies of pathogens

It is often assumed that resistance against biocontrol does not exist. Why this is not true is clearly explained in an excellent review by Duffy et al.⁹⁰. Here, we will summarize the known mechanisms of pathogen defense against biocontrol agents. They are illustrated in Figure 2, in which the numbers in brackets refer here, and in the figure, to the various defense mechanisms.

Figure 2. Major mechanisms used by microbes to defend themselves against antibiotic attack by other microbes.

Microbe A (top) produces antibiotic molecules (red squares), which enter the antibiotic-sensitive Microbe B (bottom) and fit snugly in their molecular target, thereby inactivating Microbe B (see left part of figure). A variety of mechanisms can be used by Microbe B to defend itself. The numbers in brackets refer here, and in the figure, to the various defense mechanisms. {1} Microbe B can secrete a secondary metabolite (SM), which represses the expression of the antibiotic biosynthetic genes at the level of their promoter. In many cases, acyl homoserine lactones (AHLs) are required for activating antibiotic synthesis (see top right). In such a case, Microbe B can develop one of two different mechanisms. Firstly, it can secrete an SM which inhibits the synthesis of the AHL {2}. Secondly, Microbe B can secrete an enzyme which inactivates the AHL {3}. Also, after the antibiotic has entered Microbe B, B can develop various defense mechanisms. Firstly, it can produce an enzyme which inactivates the antibiotic {4}. Secondly, the antibiotic can be recognized by an efflux pump, by which it is secreted from the cell before it can do much harm {5}. Finally, the receiving microbe can alter its target in such a way that the antibiotic is not recognized any longer {6}. For further explanation, see the section titled “Defense strategies of pathogens”.

Repression of weapon biosynthesis. The biocontrol strain P. fluorescens CHA0 produces Phl, which can kill the pathogen Forl. FA, a secondary metabolite secreted by Forl, appeared to repress Phl gene expression at the level of the phlA promoter⁹¹ (Figure 2 {1}).

Van Rij et al.⁹² have shown that, in P. chlororaphis strain PCL1391, FA also represses the production of another antibiotic, PCN, but in a different way. Like in strain CHA0, antibiotic synthesis in PCL1391 requires activation of the quorum-sensing regulatory genes phzR and phzI by the AHL N-hexanoyl-L-homoserine lactone, which, like most AHLs, can diffuse freely through the membrane into the medium. The phz biosynthetic genes are activated only once the intracellular AHL level reaches a certain threshold concentration or quorum. This phenomenon is called quorum sensing. Van Rij et al.⁹² discovered that FA in PCL1391 acts through inhibition of the synthesis of the AHL N-hexanoyl-L-homoserine lactone (Figure 2 {2}). The reduction of the PCN level appears to be the result of direct or indirect suppression of the quorum-sensing regulatory genes phzR and phzI⁹². This can be considered a form of signal interference (see the following section). Apparently, two different mechanisms have evolved for the inhibition of antibiotic synthesis in pseudomonads by Forl through FA secretion.

Signal interference. Many bacteria produce antibiotics and other pathogenicity and virulence factors, such as fungal cell wall-degrading enzymes, only at a high bacterial cell density. The induction occurs when the level of quorum-sensing molecules such as AHLs reaches a certain threshold concentration in the growth medium (see 93,94 and previous section).

Signal interference is a biocontrol mechanism based on the degradation of the AHL⁹⁵ (Figure 2 {3}). Enzymes which are able to degrade AHLs are (i) AHL lactonases, e.g. of Bacillus thuringiensis strains, which hydrolyze the lactone ring⁹⁵, and (ii) AHL acylases, which break the amide link⁹⁶.

Interestingly, it was shown recently that AHLs also play a role in the formation of biofilms. AHL-degrading enzymes inhibit biofilm formation^94,96, thereby preventing the microbes from becoming more tolerant to antibiotics.

Detoxification of the weapon (Figure 2 {4}). (i) Hydrogen cyanide (HCN) is produced by many biocontrol agents, especially pseudomonads. The cyanide ion is a potent inhibitor of many metalloenzymes, especially copper-containing cytochrome c oxidases. This volatile antibiotic shows little selectivity towards fungi⁶. Osbourn⁹⁷ has reported that several fungi are able to detoxify HCN by degrading it to formamide.

(ii) Schouten et al.⁹⁸ screened a collection of 76 plant-pathogenic and 41 saprophytic F. oxysporum strains for sensitivity to Phl; 20 of these were relatively tolerant to high Phl concentrations, 18 of these were capable of metabolizing Phl, and, for two of these tolerant strains, it was shown that deacetylation of Phl to the less fungitoxic derivatives monoacetylphloroglucinol and phloroglucinol is (part of) the mechanism of Phl degradation.

Active efflux of the weapon (Figure 2 {5}). Schoonbeek et al.⁹⁹ reported that the expression of several ABC transporter genes of the fungal pathogen Botrytis cinerea is induced by the antibiotics Phl, PCA, and PCN. Phenazines strongly induce the expression of BcatrB. The BcatrB gene encodes one of the ABC transporters. BcatrB mutants are significantly more sensitive to phenazines than are their parental strain. Phenazine-producing strains of Pseudomonas are more strongly antagonistic in vitro to BcatrB mutants than to the parental B. cinerea strain. Pseudomonas strains that produce phenazines more effectively reduce gray mold symptoms induced by a BcatrB mutant than by the parental strain on tomato leaves. It can be concluded that it is likely that phenazines are secreted by the BcatrB protein⁹⁹.

Modification of the target (Figure 2 {6}). Osbourn et al.¹⁰⁰ observed that a range of pathogenic fungi can tolerate HCN and that, in some of these fungi, HCN tolerance is due to cyanide-resistant respiration.

Dual role of fusaric acid in the war between the biocontrol bacterium P. chlororaphis strain PCL1391 and the fungal pathogen Fusarium oxysporum f. sp. radicis-lycopersici. One of the best-understood examples of attack and defense between microbes is the interaction between the fungus Forl and the bacterium P. chlororaphis strain PCL1391. The fungus secretes the secondary metabolite FA, which acts as a chemoattractant for the bacterium. This results in a high density of bacteria in the neighborhood of the fungal hyphae, in colonization of the hyphae, and eventually in biofilm formation on the hyphae. These high bacterial concentrations are ideal for quorum sensing and result in the production of the antibiotic PCN and subsequent killing of the fungus.

However, it appears that FA also has another role, namely in defending the fungus by inhibiting AHL production, and consequently also PCN production, by the bacterium. Based on this knowledge, one would predict that the bacterium can win only when a sufficiently high level of PCN is produced in a short period of time. This requires a high bacterial density. In practice, this can be reached, since P. chlororaphis strain PCL1391 is a good biocontrol bacterium¹⁸.

Is tolerance towards biocontrol agents an emerging threat? It is widely assumed, and often reported, that biocontrol strains use multiple mechanisms of attack. Since simultaneous tolerance development towards more than one mechanism in the same pathogen cell is unlikely, pathogen targets of biocontrol agents will suffer less from tolerance than organisms attacked by a single toxic compound. During the development of biocontrol agents, it would therefore be wise to select strains which use at least two mechanisms of activity against pathogens, also under the conditions where they will be used. Equally, it would be worthwhile to determine, in advance, the mechanisms of resistance to these agents and if these induce cross-resistance to diverse mechanisms of attack.

An alternative approach to reduce potential resistance would be to apply a combination of strains with different mechanisms of biocontrol. However, our experience with many such “cocktails” was that the biocontrol results were never better than that of a single strain and sometimes even worse (F. Kamilova and B. Lugtenberg, unpublished data). This negative result may be due to the limited “carrying capacity” of the root system for microbes. This would result in a dilution of the cell numbers of the beneficial strains on the root and therefore in a reduced biocontrol efficacy.