Dangers of assuming the effectors or mechanisms are known
While plant growth promotion by bacteria can be associated with the mechanisms described earlier, it would be unwise to assume that these are the only mechanisms at play, particularly under field conditions. In antibiotic research, it is assumed that we have scratched only the surface of identifying antibiotics produced in nature because of the challenge related to recreating the conditions required for antibiotic production in the laboratory (Begani et al., 2018). The same may very well be true for PGPR: the growing conditions, environmental stimuli and media composition may not lead to the activation of mechanisms responsible for plant growth promotion in natural settings. It is likely that there are mechanisms for microbe-associated plant growth promotion that we currently do not understand, and we should keep this in mind as we continue to explore these complex interactions. For example, laboratory screening is limited because it often considers only one or a few microbes inoculated onto plants, while in a field setting, the interactions among microbes are vastly more complicated. This section describes how signals exchanged between plants and microbes, in the form of root exudates and secondary metabolites, and among microbes play a role in establishing plant- microbe relationships. While this is one frontier that remains to be elucidated, there are likely to be others that have not yet even been discovered.
Root exudates are secreted by plant roots and contain soluble organic compounds, mainly including sugars, amino acid, organic acids and enzymes (Koo et al., 2005). Those compounds have important roles both in the solubilization and the mobilization of minerals and micronutrients (Koo et al„ 2005). In addition, root exudates have strong impacts on the growth of plant-associated microbes (members of the phytomicrobiome), including PGPR (Beauregard, 2015), as root exudates constitute the main reduced- carbon source for PGPR. The composition and level of root exudation can vary depending on plant growth conditions, both biotic and abiotic stress. For instance, low salinity concentrations stimulate roots to excrete more organic and amino acids, a change in exudate quality, while nutrient deficiencies could increase the overall quantity of root exudate (Koo et al., 2005). In Macrophylla, a salt-tolerant plant species, the concentration of salicylates, proline and/ or hydroxyl-L-proline in root exudates is higher than that in root exudates produced by Carrizo (Vives-Peris et al„ 2018). Moreover, the presence of microbes can, through release of specific secondary metabolites, affect root exudation, for instance, with regard to amino acid composition (Phillips et al., 2004). The root exudates produced by specific plant species probably have a range of functions. Li et al. (2016) showed that root exudates of maize can stimulate faba bean to fix N2, while this was not found to be the case in barley and wheat.The variation of root exudates may interact with specific PGPR, which enables them to colonize the plant roots, and exerts control over which types become prominent in the phytomicrobiome (Czarnota et al., 2003; Rudrappa et al., 2008; Chaparro et al., 2013). Vives-Peris et al. (2018) concluded that root exudates collected from C. macrophylla can promote rhizobacteria P. putida KT2440 and Novosphingobium sp. HR1a growth under salt stress (60 or 90 mM NaCI) and heat stress (40°C) conditions.
Data collected from microbial ecology studies has shown that there is a core root microbiome that is influenced first by soil type and second by plant species (Yeoh et al., 2017). Plants can select for desirable microbes in the rhizosphere by secreting root exudates including sugars, poly-alcohols, organic acids, fatty acids and polycosanols (lannucci et al., 2017). The composition of root exudates has been influenced by the domestication of crop species, and there is a high degree of heritability that influences the composition of root exudate profiles in, for instance, tetraploid wheat (lannucci et al., 2017). There is a potential to engineer the plant production of particular root exudates to recruit beneficial microbes (Mohanram and Kumar, 2019). Root exudates can also induce antibiotic production by beneficial microbes in the rhizosphere, which allows for the successful colonization by beneficial microbes (Ogran et al., 2019).
Interspecific and intraspecific microbial interactions in the rhizosphere
Communication among microbes, not just between plant and microbes, also contributes to the establishment and maintenance of plant-microbe interactions. However, our current culture-dependent approach has focussed on the application of individual PGPR strains. Furthermore, plant growth-promoting microbial research is biased towards microbes that grow quickly in pure culture on artificial, high-nutrient media (Nai and Meyer, 2018). This limitation has been discussed in the context of antimicrobial discovery, and Molloy and
Hertweck (2017) proposed the adoption of an ecological paradigm to discover antimicrobials. This strategy could also be beneficial in the search for microbe- to-plant signal molecules. It seems it is now time to investigate how relationships between microbial species in the rhizosphere impactthe production of microbe- to-plant signal molecules and, in turn, how the production of microbial secondary metabolites influences plant growth under optimal and stressed conditions.
Intermicrobial signalling affects on microbial metabolism
Bacteria can produce a large array of bioactive molecules, and we have only begun to scratch the surface when it comes to the identification and isolation of molecules for industrial uses (Berdy, 2012). Because bioactive molecules are generally secondary metabolites, which are energetically expensive to produce, they are not constitutively expressed. In fact, recent research has shown that the genomes of many bacteria and fungi contain code for gene clusters that are expressed only after being activated - these are referred to as silent or cryptic gene clusters. The observation of this phenomenon led to the development of the 'one strain many compounds' (OSMAC) hypothesis, which states that one bacterial strain has the ability to produce different bioactive molecules under different growing/fermentation conditions (media composition, e.g., C, phosphate, N sources; aeration rate; temperature; pH; inclusion of inhibitors or growth enhancers) (Bode et al., 2002).
Multiple reports have promoted a paradigm shift from classical microbiology, where strains are cultivated in pure culture in favour of mixed cultivation containing two or more microbial strains (Nai and Meyer, 2018). The advantages of mixed cultures are that the consortia of microbes are able to perform more complex tasks compared to pure cultures (Brenner et al., 2008; Shank and Kolter, 2009; Bertrand et al., 2014). Within mixed cultures, members of the community are constantly excreting metabolites into the culture. These metabolites serve as intermicrobial signals, which can activate otherwise-silent gene clusters, resulting in the production of a larger array of antibiotics (Netzker et al., 2018) to allow for competition between species or can result in cooperation between species (Garcia et al„ 2016). This can result from obligatorily mutualistic metabolism, also known as metabolic syntrophism (Stubbendieck et al., 2016). These interactions allow microbes to work together to access nutrients available from the environment that they would not be able to access via their own metabolic pathways while maintaining smaller genome sizes (Stubbendieck et al., 2016).
Several studies have been conducted on bacterial-bacterial co-cultures. For example, the co-culture of two bacteria, Staphylococcus lentus SZ2 with Vibrio harveyi, an aquaculture pathogen, led to increased biosurfactant production by S. lentus and reduced growth of V. harveyi planktonic cells and biofilms (Hamza et al., 2018). The authors hypothesized that the change in the growth rate of V. harveyi was related to the competition from S. lentus. B. subtilis was also shown to produce bacillaene, which inhibits the growth and pigment production of several Streptomyces spp. and contributes to defence against the soil-dwelling bacterium Myxococcus xanthus (Stubbendieck et al., 2016). This highlights the diverse effects of a single bacterial secondary metabolite in relation to other soil microbes.
There are also several examples in the literature of mixed bacterial-fungal cultures producing bioactive secondary metabolites that are not produced in pure culture (Rateb et al., 2013). For example, cocultures of Sarocladium strictum and Fusarium oxysporum led to de novo production of the toxin fusaric acid, which was involved in virulence associated with fungal infections (Bohni et al., 2016); furthermore, this effect was strain-dependent. Zhang et al. (2017) identified new molecules with potent a-glucosidase inhibitory activity produced by the co-culture of Trichoderma sp. 307 and Acinetobacter johnsonii B2; these molecules could be useful in the treatment of type 2 diabetes. Another study found that the co-culture of Bacillus amyloliquefaciens and Setophoma terrestris resulted in the production of a novel blennolide that had promising and unique effects on prostate cancer cells (Arora et al., 2018). In the previous three examples, researchers proposed that the bacteria induced bioactive molecule production bythefungal species, butdid not otherwise investigate the underlying mechanism of de novo bioactive molecule synthesis. Netzker et al. (2015) provided a recent summary of our current understanding of mechanisms underlying de novo bioactive molecule biosynthesis by fungi in co-culture with other microbes, including a case study of the interaction between Aspergillus and Streptomyces rapamycinicus, and a discussion of the role of chromatin remodelling in microbial interactions. The authors highlighted the need for an improved understanding of the mechanisms underlying the biosynthesis of secondary microbial metabolites in order to benefit from the production of these molecules in the context of therapeutic agents; the same can be said for benefiting from the use of these molecules in agricultural systems.
Other forms of microbial interactions
Microbes use not only chemical signals to communicate, cooperate or compete but also physical signals. For example, in anaerobic methane-producing or methane-consuming communities, microbes can communicate via electron transfer over relatively long distances (pm to cm)(Lovley, 2016).This mechanism may also contribute to gut microbiome communication with host gut cells and has been used to stabilize anaerobic digestion. Most importantly, for plant scientists, this interaction has been demonstrated to occur in rice paddy soil and sediments and Arctic peat (Lovley, 2016).
To date, there are several examples of how microbial consortia can be applied in other microbiological fields. For instance, bacterial-fungal interactions have been harnessed for natural product discovery and could also prove to be useful in plant growth promotion (Haq and van Elsas, 2015). As we begin to understand how microbial consortia achieve community-level properties in natural rhizosphere environments, we can begin to use synthetic microbial ecology to achieve desirable community-level properties for agricultural applications (Dolinsek et al., 2016). Specifically, consortia could be designed to produce molecules that promote plant growth by reducing the build-up of inhibitory compounds (Santala et al., 2014; Jones and Wang, 2018). This will become easier as we begin to understand how to define the function rather than the taxonomic composition of the plant microbiome (Deines and Bosch, 2016). In addition to using microbial consortia with specific functions for plant growth promotion, there is potential to apply functional microbial consortia for bioremediation, energy production and therapeutics (Hays et al., 2015).
Bioactive molecules produced by microbes can have influences on microbial metabolism (within or between microbial species) either directly or indirectly on plants. For example, thuricin 17, a small peptide produced by B. thuringiensis NEB17,and lipochitooligosaccharide, produced by Bradyrhizobium japonicum, improved soybean germination under salt stress conditions up to 150 mM by enhancing PEP carboxylase, rubisco oxygenase large subunit, pyruvate kinase and isocitrate lyase (Subramanian et al., 2016a,b). In contrast, a molecule secreted by Pseudomonas sp. 23s shows in vitro antagonistic activity against the serious tomato pathogen Clavibacter michiganensis subspecies michiganensis (Yoko Takishita, personal communication). Inoculation of tomato seedlings with the same strain stimulates growth and induces systemic resistance (Takishita et al., 2018), but it remains to be determined whether this effect is associated with the same molecule that shows biocontrol activity. Only a small number of studies have reported the effects of microbial signal molecules on plant growth, and it is likely that we have identified and/or characterized only a tiny fraction of the molecules responsible for microbial effects on plant growth. Zachow et al. (2015) demonstrated that a molecule secreted by an endophytic Pseudomonas is involved in the suppression of a pathogenic Rhizoctonia solani and root colonization, which highlights that these molecules can be involved in multiple functions (Zachow et al., 2015).