Mechanisms: Modes of Action

There are four main modes of action underlying biological control of plant diseases: (a) exploitation competition for resources (oxygen, carbon, nitrogen and other essential resources); (b) interference competition for space via antibiosis where the BCA inhibits the pathogen via effects of toxic secondary (specialized) metabolites; (c) hyperparasitism, where the antagonist acts as a predator and exploits the pathogen as a prey; (d) induced resistance—the indirect interaction of a BCA via induction of plant defence mechanisms against invading pathogens. A fifth mechanism that can contribute to disease control is plant growth stimulation by better nutrient absorption and/or by affecting plant hormone pathways, as demonstrated by, for example, various rhizosphere bacteria and fungi. A strongly growing plant may be better able to withstand a pathogen and a rapid establishment of seedlings in the field can lead to avoidance of damping-off diseases. However, some researchers would not consider this as biological control.

A single BCA may exhibit a combination of these modes of action. The individual modes of action have different but not exclusive population dynamic consequences. It can be quite difficult to prove that a particular mechanism is operating in planta, even though it can operate in vitro. More than one of these mechanisms can contribute to a concerted action in a particular case and the importance of a specific mechanism used can vary from case to case, even using the same organism. For example, species of Trichoderma and Clonostachys may act as hyperparasites, metabolite producers, competitors and/or modulators of plant defence responses. Exploitation competition can be independent of the pathogen population size, simply reflecting efficient local resource capture. Competition through more efficient resource use does not rely on direct interaction as the BCA takes over resources and space meaning the pathogen cannot thrive. Being the first to colonize new resources is another important type of exploitation competition that can deprive a pathogen of the resources it requires, especially in the critical early stages of colonization. In addition, the ability of beneficial organisms to colonize a substrate that is not preferred by the targeted pathogens could improve competitiveness of the biocontrol agent against the pathogen in the targeted pathogen community. Alternatively, interference competition through antibiosis, depending on how close the organisms need to be to interact, may allow the BCA to monopolize the habitat. Hyperparasitism requires that the BCA occurs and is metabolically active spatially close to the target pathogen (normally in the niche where the pathogen would infect, or which is occupied by fruiting bodies or resting structures of pathogens that are parasitized by a BCA).

The question has been raised whether pathogens could evolve to be resistant to BCAs, as frequently occurs with repeated use of pesticides with specific modes of action. Over more than four decades of using biological control, resistance in the target bacterial and fungal pathogens has yet to be demonstrated. The direct use of metabolites and extracts, which is not included in the definition of biocontrol discussed earlier, leads to high exposure of the pathogen; this may carry more risk of development of pathogen resistance, as seen with the regular use of chemical pesticides. In the case of bacteriophages, it is known that bacteria can adapt rapidly to bacteriophages and are expected to overcome single strains. Products are therefore based on cocktails of bacteriophages to reduce this problem (see below).

Although resistance has not been considered a serious problem for most other practical uses of biocontrol, we will next discuss the issue and its relation to mode of action. It is not easy to see how a pathogen could evolve resistance to exploitation competition in nature. However, as for chemical pesticides, resistance towards BCA metabolites in pathogen populations is a theoretical possibility; if a substantial proportion of a pathogen population is regularly exposed to this metabolite, leading to high selection pressure, resistant phenotypes could, in principle, arise. Some BCAs mainly rely on antibiosis via the production of one or a few specific toxic metabolites; resistant pathogen phenotypes could result, perhaps with a consequent risk of field resistance. An example of one stage in this process has been observed in take-all decline of wheat, induced by monoculture, in suppressive soils. Isolates of the pathogen involved (Gaeumannomyces tritici) showed variation in sensitivity to two metabolites produced by strains of Pseudomonas fluorescens that were claimed to be important for disease suppressiveness. Such variation in different traits is to be expected but there is no clear evidence that the population as a whole has become less sensitive to the two metabolites tested (phenazine-1-carboxylic acid or 2,4-diacetylphloroglucinol) despite heavy exposure to them. Furthermore, no evidence of resistance to 2,4-diacetylphloroglucinol was found in of G. tritici populations from Washington State, USA.

In general, pathogenic organisms can be expected to vary in traits allowing them to thrive in variable but competitive environments. Because resistance to a metabolite can be conferred by changes in the target site, detoxification, excretion (efflux) or general metabolic adjustments, intensive use of a BCA acting via antibiosis and based on one or a few specific toxic compounds could lead to the evolution of resistant pathogens. The selection pressure is increased if pathogen populations experience heavy (long term and/or highly effective single dose) exposure to the metabolite. For this reason, vulnerability to resistance should be considered on a case-by-case basis when creating strategies for biocontrol use. There is a strong argument for the development of many different BCAs for a given problem, to avoid exposure of large proportions of the pathogen population to the same selection pressure.

A special case where a strategy for avoiding resistance in pathogen populations has been addressed is the biocontrol of crown gall, caused by the bacterium A. tumefaciens, using the BCA A. radiobacter (syn. R. rhizogenes) strain K84 that produces the toxin agrocin responsible for the antibiosis. Here, the BCA harbours a plasmid that encodes resistance to its own agrocin toxin and at the same time encodes mobility of this plasmid with resistance to other Agrobacterium strains. In this case, the concern was that the plasmid might be transferred to the plant-pathogenic Agrobacterium bacterial strains making them resistant to agrocin. As this was demonstrated to happen both in field and in laboratory experiments and information accumulated that it also might be happening under commercial use, a gene modification of the BCA was created in which the plasmid mobility trait was deleted—strain K1026. This strain K1026 has been used commercially in Australia and in the United States, although biocontrol with commercial use of the wild-type strain K84 still provides effective biocontrol in many crops worldwide after almost 50 years of commercial use. Strict legislation for regulating BCAs has until now prevented the use of R. rhizogenes for biocontrol in the EU, but both the mutant K1026 and the wild-type K84 are approved in many other countries.

A specific (biotrophic) hyperparasite requires a pre-existing host population to parasitize as well as a living host for activity and growth; therefore, they will be effective in the short term only if applied inundatively. An exception may occur if a biotrophic hyperparasite could function effectively and survive in the longer term in an environmental reservoir. Unfortunately, such biotrophic hyperparasites will not usually compete well in the absence of a host. Nonetheless, there are some examples of commercialized biotrophic hyperparasites used for biocontrol such as Ampelomyces quisqualis, used against powdery mildew and Coniothyrium minitans, a parasite of several sclerotia-forming plant pathogens. A special example of a potential BCA is the hyperbiotrophic fungus Pseudozyma flocculosa (a yeast) that parasitizes powdery mildew and, in this way, obtains access to resources from the leaf infected by the mildew fungus. P. flocculosa is dependent on a living host–pathogen combination and so needs to find a new host mildew if a mildew colony dies. Interestingly, P. flocculosa also produces an antifungal glycolipid, flocculosin, suspected to have a role in the interaction. However, a CRISPR-Cas9 mutant impaired in the biosynthesis of flocculosin was apparently unaffected in its ability to antagonize powdery mildew. This is an effective lifecycle as powdery mildews are polycyclic, the organism attacks multiple species of powdery mildew, and new infections are found throughout the growing season in many crops, continually offering new living hosts for the BCA.

Whether the use of specialized hyperparasitic BCAs would be risky in an inundative strategy should be considered case by case. It is possible to set up an effective strategy for their use, provided knowledge of the target pathogens and their disease cycles, the environmental conditions, the biology and ecology of the BCA allow the prediction of the right timing and placement of the BCA in the niches where it is to act. A. quisqualis, for example, is effective against powdery mildew on cucumber but less effective in controlling powdery mildew on grapevine caused by Uncinula necator, as it mainly parasitizes the fruiting bodies (chasmothecia) late in the season. Parasitism of the conidial stage throughout the growing season is highly dependent on humidity, which is not a requirement for conidial production by the pathogen. Therefore, the BCA is less efficient in periods with low rainfall/humidity. However, parasitism of chasmothecia might have an important role in integrated disease control by reducing primary inoculum for the following year.

Although not a crop example, the rust hyperparasite Sphaerellopsis filum appears to have specific genotypes that are adapted to attack only some genotypes of individual species of grass-infecting rusts; this phenomenon might also be relevant to other biotrophic hyperparasites. Viruses can also be considered as obligate hyperparasites with more or less specific host ranges.

However, most BCAs that work via hyperparasitism are necrotrophic parasitic fungi that compete well and survive without a living host pathogen. Examples are species of Trichoderma and Clonostachys that can operate as mycoparasites as part of their lifestyle but also grow and multiply via other ways of life. Necrotrophic hyperparasites are considered more aggressive as BCAs than the more specialized hyperparasites, and are more competitive, for example, in the rhizosphere and in root colonization.

Induced resistance is a well-studied phenomenon in the laboratory where there are good examples of this as a mode of action. However, induced resistance will be ineffective against existing high population densities of a pathogen. Interaction with target pathogens via induced resistance does not require close proximity of the target and the BCA. For example, root application of S. indica can stimulate both plant growth and induced resistance in the shoot. Volatile specialized metabolites can act as signals between plant parts and, at least in principle, between neighbouring plants. Moreover, application of BCAs can induce resistance in the progeny of treated plants, a phenomenon termed transgenerational systemic acquired resistance. Several phytohormones have been shown to be involved in the resistance induced by S. indica. Hormones have complex and sometimes antagonistic effects, which can influence the response to both abiotic and biotic stress, modifying cellular physiology to respond and adapt to the stress. For pathogens, the activated defence responses provide induced resistance.

Understanding the evolutionary response to the use of host resistance inducers raises the question of why plants do not trigger these defences constitutively. The obvious answer is that induced resistance requires energy or involves intrinsic damage such as cell death, and so there is a fitness cost. This means that the induced defences are regulated (e.g., by transcriptional modulators such as NPR1) and not deployed unless needed. In this case, therefore, the use of a BCA to induce resistance in the absence of a substantial subsequent pathogen attack could lead to loss of yield. This would be a serious set-back in developing BCAs as part of an integrated disease management toolbox. Although negative effects of application in the absence of pathogens are hard to demonstrate, experiments involving transgenic plants where constitutive expression of R genes and regulators such as Npr1 were used can result in enhanced induced resistance with demonstrable fitness costs. One of the great challenges for the genetically modified organism (GMO) approach in recent decades has been the identification of appropriate promoters for driving the expression of such genes. The use of tissue-specific promoters can mitigate the negative effects of inappropriate expression.

The effect of S. indica (and some other agents) has been suggested to be first and foremost growth promotion allied to effects such as drought tolerance. In that case, the question remains of what prevents evolution of constitutive expression of the growth-promoting traits. Apparently, defences can be, if not activated, primed, with effects on growth and yield that are too slight to measure. There are two hypotheses that could explain why the defences remain facultative: (a) the costs are expressed in specific circumstances, not usually encountered in experimental or field-crop settings; (b) less probably, it may be that in natural settings, with a diverse and microbe-rich soil, priming always happens, so there is no selective advantage or disadvantage in facultative control—it is just a normal stage in development. If (a) is correct, there is the important practical conclusion that we should be looking very hard for side-effects of these priming organisms before they are too widely deployed on crops.