The modulating effect of bacterial volatiles on plant growth

Introduction

Bacteria are smelly—this is common knowledge to most of us, yet the fact that plants are also able to “smell “ them, and to react to these odours by increased or reduced growth, is a relatively recent discovery. In 2003, Ryu and co-workers described this phenomenon for the first time¹ and opened the door to a new field of research: the volatile-mediated impact of bacteria on plants. Since then, contrasting effects have been reported, ranging from large plant biomass increase to plant death.^1-6 Despite the strong effects observed, the elucidation of the responsible molecules is not as advanced as one might expect; very few bacterial volatile organic compounds (VOCs) have been unequivocally identified as responsible for the observed positive or negative impact on plant growth (Fig. 1). This might be due to the high complexity of the volatile blends produced by bacteria. More than 300 candidate molecules have been identified thus far,⁷ yet GC-MS analyses often reveal a high proportion of unknown compounds, suggesting a great potential for the discovery of new metabolites. While most research has been performed using simple, divided Petri-dish experimental setups and rich culture media for bacterial growth, there are at least two reasons to believe that volatiles might play an important signalling role between bacteria and plants in the natural situation as well. The first is that volatiles are known to be used by plants as warning or attraction signals in plant-plant⁸ and plant–insect⁹ communication, suggesting that recognition mechanisms exist which enable plants to detect volatiles. The second reason concerns the rhizosphere environment, which seems favourable for volatile-mediated communication, since the partners are spatially close to each other and volatiles are more likely to accumulate and reach their activity threshold in the rhizosphere than in the wind-exposed above-ground environment (Fig. 1). Eight years after the first report of plant growth promotion by bacterial volatiles, the phenomenon has been confirmed by independent studies using different strains and plant models, yet the molecules responsible for the observed effects, as well as the metabolic changes occurring in plants upon detection of bacterial volatiles are still poorly understood. The aim of this review is to summarize the current knowledge of the volatile-mediated effect of bacteria on plants, to highlight gaps which prevent us from practical use of promising effects and to point out future lines of research which should help us to better understand this fascinating new facet of bacterial–plant communication.

Contrasting Effects of Bacterial Volatiles on Plant Growth

Most of the studies investigating the volatile-mediated impact of bacteria on plants have been performed using a divided Petri dish setup, where two compartments are separated by a plastic border. This allows volatile compounds to be exchanged, while preventing any diffusion of non-volatile metabolites through the medium. Apart from a few reports on other model plants,^3,10-12 most of the available data originates from assays using Arabidopsis thaliana as a target for bacterial volatiles (see Table 1). A. thaliana is usually grown on MS (Murashige and Skoog) medium supplemented with sucrose and in most cases, pre-germinated seedlings are transferred to a divided Petri dish to be exposed to bacterial volatiles. When considering the effects reported so far (summarized in Table 1), it appears that growth inhibition was much more frequently observed than growth promotion. A closer look into the culture conditions reveals that promoting effects, as described mainly for Bacillus species, were only observed when MS was used to grow the bacteria as well as the plants.^1,13,14 When LB (or the similar medium NA) was used instead, the volatile-mediated effect of bacteria was in almost all cases deleterious (Table 1). MS and LB differ greatly in their composition: MS is a mineral medium with sucrose as the C source, it has an acidic pH and low agar concentration while LB is a slightly alkaline, complex medium containing mainly hydrolyzed proteins and higher agar concentration. It is thus not surprising that the same bacteria grown on LB and MS should produce varying blends of volatiles with differing effects on the plants. Moreover, the growth kinetics of the same strain will differ between LB- and MS-based growth, with much quicker growth on LB than on MS. Therefore it would be expected that when plant seedlings at the same development stage are tested with bacteria on both media, the peak of volatiles production (which is expected to occur during bacterial stationary phase¹⁷) would occur at different stages in the development of the plants. One might postulate that very young plants will be more susceptible to potentially harmful volatiles than older plants. This might account, at least partly, for the more negative response of plants toward bacterial volatiles produced on LB vs. MS. In an experiment using M. sativa as a model plant, Zou et al. assessed the plant response to the volatiles of A. agilis supplied at different time points after germination.¹⁴ Although they did not see any negative effects using NA as a culture medium, which could be related to the use of a different, larger and likely more resistant model plant than A. thaliana, the effects observed on the M. sativa were influenced greatly by the time point of bacterial inoculation. This suggests that in addition to the blend of volatiles produced, the time at which volatiles are supplied to the plant is a factor that significantly influences the outcome of the interaction. In order to assess the impact of the cultivation medium on the volatile-mediated effect of bacteria on plants, Blom and coworkers tested a collection of rhizosphere strains, mainly belonging to the Burkholderia genus, and grew them on four different media. This experiment demonstrated that the same strain grown on different media can have a range of effects on plant growth.¹⁸ Killing effects were only observed with LB, while growth promotion was obtained on all media tested, including two rich complex media, MS as well as the nutrient-poor, soil mimicking Angle medium.¹⁹ The outcome of this screen was that of approximately 40 strains, every single one caused a significant impact on A. thaliana in at least one of the culture media used.¹⁸ Taken together with the available literature (summarized in Table 1), this suggests that volatile-mediated plant growth modulation by bacteria is a widespread mechanism.

The Search for Active Volatiles

In addition to the production of CO₂, which is likely to moderately contribute to the growth promotion observed in closed Petri dish assays,^20,21 other compounds must be responsible for the large increase in plant biomass reported as occurring upon exposure to bacterial volatiles.^1,6,18,22 In the first report of volatile-mediated effects of bacteria on plants, 2,3-butanediol was put forward as the main causative agent of the observed effects, based on the use of butanediol-deficient B. subtilis mutants and on the application of the pure compound.¹ These results were confirmed by later studies using P. chlororaphis, which indicated that butanediol was not only promoting growth, but inducing systemic resistance and mediating drought tolerance.^11,16 Butanediol, the end product of a specific fermentation pathway which is used by certain bacteria to avoid acidification, is likely to be produced on the sucrose containing, low pH MS medium used in the above mentioned studies.¹ The medium MR-VP (Methyl-Red-Voges-Proskauer) favors butanediol fermentation and is therefore routinely used to assess whether bacteria are able to carry out this specific fermentation. Interestingly, Blom and coworkers observed that most Burkholderia strains tested promoted plant growth when cultured on MR-VP medium, although this is not associated with butanediol fermentation, as this pathway is not present in members of this genus.¹⁸ This suggests that butanediol is not the only candidate to explain the large growth promotion observed. In addition to butanediol, only two other compounds have been shown to promote plant growth when applied as pure substances: dimethylhexadecylamine, which increased the biomass of M. sativa by a factor of about 1.4,²³ and 2-pentylfuran which increased the biomass of A. thaliana by a factor of about 2¹⁰. A number of molecules responsible for the observed volatile-mediated deleterious effects on plants have been described. These caused quantitatively comparable effects with the exposure to the complex blend of bacterial strains. Early observations led to the hypothesis that hydrogen cyanide, which is produced by a small number of bacterial species including some Pseudomonas and Chromobacterium species,²⁴ might account for the negative impact of rhizosphere pseudomonads on wheat.¹² Later studies on A. thaliana confirmed this hypothesis using cyanogenic bacterial strains, the corresponding null mutants, as well as exposure to pure hydrogen cyanide.^6,15 Serratia species, which produce no or negligible cyanide,⁶ have also been observed in several studies to inhibit A. thaliana. These negative effects were ascribed to dimethyl disulfide and to the inorganic volatile NH₃.¹⁷ While these results are interesting and constitute a promising beginning in the identification of active molecules, they do not explain the full extent of the effects observed upon exposure to bacterial volatiles, especially where promoting volatiles are concerned. It is likely that the mixture of compounds, rather than single molecules, account for the observed effects, and this should be considered in future studies.

First Insight into Signaling and Metabolic Changes

Colonization of the root system by plant-growth promoting rhizobacteria (PGPR) has been suggested to stimulate host plant growth by synthesis and secretion of the canonical phytohormones auxins, cytokinins and gibberellins or by the breakdown of ethylene via 1-aminocyclopropane-1-carboxylate deaminase.^25-28 Other nutrient-oriented strategies have been described, where bacteria increase N, P and Fe availability in soil.^29,30 However, recent studies have clearly demonstrated that, in the absence of physical contact with the plant root, biologically active bacterial volatiles with no relationship to classical plant hormones can likewise trigger plant growth promotion.^1,30

Although VOCs-mediated plant–bacteria interactions have not been extensively documented, it appears that the variety of active VOCs increases with the number of strains investigated.^{1,13,18,31,32} In contrast, the plant panel of responses to these bacterial signals is restricted to the physiological pathways involved in growth control. Thus far, the individual contribution of these pathways to the observed growth alterations remains elusive. Previous studies focused on the involvement of four major hormones in the transduction of the bacterial signals: ethylene, cytokinins, abscisic acid and auxins.^1,13,33-35 The putative impact of other hormonal effectors such as gibberellins, jasmonic acid and salicylic acid has also been implicated^1,13 but there is too little evidence to allow a full analysis. We will thus focus on ethylene, cytokinins, abscisic acid and auxins in this review.

Future Challenges

The available literature on volatile-mediated bacterial–plant interactions suggests that the production of plant promoting volatiles is a widespread feature among rhizosphere bacteria.^1,7,18,30 The wide variety of bacterial strains and plant species reported to engage in such VOCs-based interactions suggests that physiological responses to signal molecules might be shared universally. This concept is stimulatingly supported by the fact that most of volatile bacterial compounds reported so far were already described as semiochemicals in plant–insect communication or in bacterial-bacterial and bacterial–animal interactions;² (Pherobase; SuperScent). To better understand the perception by the plant of single bacterial compounds, as well as the signaling cascades and subsequent metabolic changes they trigger will be one of the major upcoming tasks in the field. To this end, the first step is the elucidation of the active molecules and characterization of their concentration range of activity. While a few candidates have been put forward as plant-growth promoting volatiles, all extant investigations of plant responses, including the transcriptomic, proteomic and mutant-based studies mentioned above, have originated from exposure of the plants to complex blend of volatiles, rather than to single compounds. While this might be closer to the natural situation, we believe that the accurate characterization of the single active compounds and their effects on plant physiology and development is a prerequisite to better understand volatile-mediated bacterial–plant communication. Therefore, the beneficial effects on the plant of PGPR strains should be quantified in a normative experimental setup, monitoring detailed standard growth parameters. This would make it possible to investigate the action of isolated compounds, and to compare the effects obtained with the pure compounds or given mixtures with the effects observed with the strain’s volatile blend. This would potentially allow pinpointing of the network of receptors and effectors dedicated to bacterial emissions, should such a network exist in plants. A further point to clarify is the identification of the responsive plant organs. Although one would logically assume that rhizobacterial volatiles would target the root system, to our knowledge, no study has yet assessed this question, and most of the literature available relies on exposure of the plant shoots rather than the roots to bacterial volatiles.

The questions of the ecological relevance of such plant growth promotion by bacteria is as fascinating as it is difficult to tackle: the advantage for the volatile-producing bacteria is quite evident, since the fitness cost endured to produce what seems to be to a great extent metabolic waste products is likely to be largely compensated for the higher amounts of easily degraded carbon provided from increased root exudation. In contrast, the benefit for the plant is less clear in case of enhanced (or faster) growth: the competitive advantage of growing faster than other plants sharing the same soil resources seems an asset, yet the energy invested in growth is likely to be diverted from that required for protection against biotic or abiotic stresses. This might not be relevant in laboratory-grown plants but is of great importance in natural conditions.

The field of volatile-mediated interactions between bacteria and plants is still in its infancy but it holds great promise for future discoveries, of which some might be of interest for agronomy, not only due to direct growth promotion, but also to the resistance-inducing effect of volatiles, which has been repeatedly reported.^11,13,55 Once the active compounds are elucidated, experiments should be performed in conditions which are closer to the field situation than the very reductive Petri dish setup used so far in most studies. The question of whether continuous exposure of the plant to volatiles is needed for growth promotion, as suggested by,²⁴ or whether a single “whiff” of volatiles would be sufficient to trigger these effects is of relevance both for the understanding of the mechanisms and for proper field application. If the effects observed in controlled laboratory conditions could be extrapolated to the field, exposure to volatiles or direct inoculation of crops with volatile-producing bacteria might open new perspectives for the sustainable intensive agriculture of tomorrow.⁶¹