Microorganisms in the phylloplane modulate the BVOC emissions of Brassica nigra leaves

ABSTRACT

Numerous factors can affect the Biogenic Volatile Organic Compounds (BVOC) emitted by plants. One of these factors is the microbial communities living on leaf surfaces (phylloplane). Bacteria and fungi can use compounds produced and emitted by plants for their own metabolism. Thus, microorganism communities can modulate BVOC emissions and affect interactions between plants and other organisms. The aim of this study was to evaluate the role of microbial communities on BVOC emissions of Brassica nigra leaves. Therefore, we removed bacteria and/or fungi by using bactericide/fungicide treatments in a factorial design experiment with Brassica nigra grown in pots. BVOC emissions were sampled before and after the treatment application. Our results showed that four new compounds (cyclohexanone, cyclohexyl cyanide and two unknown compounds) were emitted after the removal of fungi, whereas no effect was detected in response to the bactericide treatment. This suggests that fungi inhibit or reduce the production of the above mentioned BVOCs from Brassica nigra leaves or use those compounds for their own metabolism. The origin and the roles of the novel compounds emitted requires further investigation.

Introduction

Plants emit a large variety of Biogenic Volatile Organic Compounds (BVOC) belonging to several different chemical groups. These BVOCs include terpenes (e.g. isoprene, monoterpenes and sesquiterpenes), Green Leaf Volatiles (e.g. alcohols, aldehydes and acetates) and benzenoids.¹ BVOCs have numerous different roles in plants, including (i) protecting plants against abiotic stress (e.g. water deficit, heat stress) through antioxidant properties,² (ii) attracting pollinators,³ (iii) repelling herbivores and attracting their natural enemies, and (iv) protecting against pathogens.⁴

There are numerous factors that can modify BVOC emissions, including light and temperature, atmospheric pollution,⁵ pathogen infection and herbivore attack.⁴ There is also some evidence suggesting that the communities of microorganisms living on leaf surfaces, known as the phylloplane,⁶ could alter the blend of BVOC emissions released by plants. This has been demonstrated for Sambucus nigra.⁷ Numerous studies have shown that microorganisms (fungi and bacteria) can produce and emit a diverse range of BVOCs, including fatty acids, terpenes, alcohols and aldehydes.⁸ Moreover, a relationship has been elucidated between the microorganisms within the phylloplane and their host plant, whereby the microorganisms use the different products produced and emitted by plants as substrates for their own metabolism.⁹ For instance, methanol emitted by plants during the process of vegetative growth, can be used as a carbon source by microorganisms.¹⁰ Therefore, it is likely that the quantity and/or the quality of BVOCs observed are modulated by microorganisms, though only a few studies have so far addressed this issue.⁷ It is important to understand the role of the phylloplane in the production and emission of BVOCs, since it could have a tangible role in all the biological and ecological functions attributed to BVOCs.⁶ This has also practical relevance, since in modern agriculture, phylloplane microorganisms are commonly removed via pesticides from plant leaves, which could have a significant impact on plant’s interactions with other organisms such as other plants, herbivorous insects or pollinators.

In this study, we investigated whether the removal of microorganisms from the surface of the phylloplane can change the blend and magnitude of BVOC emissions from Brassica nigra leaves. Using both a fungicide and a bactericide, we distinguished between the roles of bacteria and fungi in BVOCs observed. Our hypothesis was that the BVOC emissions would be modulated by the removal of microorganisms.

Methods and materials

Plant material

Seeds of Brassica nigra were sown in 0.8L plastic pots containing a mix of peat, soil and sand (3:1:1). Seeds were provided by E. Poelman (Wageningen University, the Netherlands). A total of 20 plants were grown in plant growth chambers (Weiss Bio 1300m, Germany) with a 16h/8h day/night cycle. Daytime conditions of 21°C and 60% relative humidity and nighttime conditions 16°C and 80% RH were maintained. Plants were watered every day and fertilized twice per week with a 0.1% solution containing nitrogen, phosphorous and potassium in a 19:4:20 ratio (Kekkilä Oyj, Finland).

Experiment

B. nigra were separated into four groups of five plants, which were each assigned to a different treatment and evaluated in a factorial design experiment: control (C), bactericide (B), fungicide (F) and fungicide + bactericide (FB). An initial round of BVOC sampling was performed 24h before the treatment application. The bactericide solution contained streptomycin sulfate (Sigma-Aldrich, USA) at 300mg.L⁻¹, the fungicide solution contained ciclopirox (Sigma-Aldrich, USA) at 50mg.L⁻¹ and the bactericide + fungicide treatment contained both products at the same concentrations as above. Water was applied as a control. Twenty mL of solution was sprayed to each plant. Twenty-four hours after the application of treatments, BVOCs were sampled for a second time.

BVOC collection and GC-MS analyses

Each plant was enclosed in a polyethylene terephthalate bag with dimensions of 35x43cm (Eskimo, Finland). Before use, each bag was heated in an oven at 120°C for 1h. Pressurized inlet air was filtered through activated charcoal and introduced into bags at 300mL.min⁻¹. After flushing the bags with clean air for 20mins, BVOC emissions were collected for 1h into pre-conditioned cartridges containing Tenax TA (Markes International, UK), which were positioned at the outlet of the bag. Cartridges were connected to a vacuum pump (D-79112, KNF, Germany) via clean silicone tubing and samples were pulled with a flow rate of 220mL.min⁻¹. Both inlet and outlet airflows were calibrated with a mini Buck calibrator (mini Buck calibrator, AP Buck, USA). After collection, cartridges were stored at 4°C before analysis.

Analysis of cartridges was performed by gas chromatography (GC, model 6890, Hewlett Packard, USA) mass spectrometry (MS, model 5973, Hewlett Packard, USA) with samples thermally desorbed with an autosampler (ATD 400 Perkin Elmer, USA). Samples were desorbed at 250°C for 10 min, cryofocused at −30°C in split mode with a ratio of 1/20. The column used to separate molecules was an HP5-MS (60m x 0.25mm x 0.5µm, Agilent, USA). The chromatography program was set up as follows: 40°C at the beginning and with a hold for 2min, 3°C.min⁻¹ temperature ramp to 210°C, and then a 5°C.min⁻¹ temperature ramp to 250°C. This final temperature was held for 5 min in order to clean the column. The carrier gas was helium.

BVOC identification was corroborated with analytical standards (Sigma-Aldrich, USA), which were injected as an external reference (and specified in Table 1 in supplementary files). Kovats indices (calculated with a C₈-C₂₀ series of alkanes) as well as NIST and WILEY libraries were used to further confirm compound identification. The analytical standards were also used to quantify the BVOC emissions.

Statistical analyses

Statistical analyses were performed with R software (v. 3.4.3). Partial Least Squares – Discriminant Analyses (PLS-DA) were performed with the package RVAideMemoire after log and auto-scaling transformation of the data. Pairwise tests were performed based on the PLS-DA, with a cross validation with 50 submodels. The graphics for the PLS-DA were built with metaboanalyst (https://www.metaboanalyst.ca/). The compounds showing the highest scores in variable importance for projection (VIP > 2) were analyzed for each treatment with a paired Wilcoxon test to highlight the before and after treatment differences.

Results

The emission rates for all individual compounds detected in the leaf blend of B. nigra are available in Table S1 in supplementary files. PLS-DA analysis of the BVOC emissions prior to application of the different treatments revealed no separation between the groups and pairwise tests did not reveal significant differences for any individual compounds (Figure 1, P > 0.05). Thus, the BVOC emissions from B. nigra leaves were the same between the different groups before the application of treatments. After application of the antimicrobial treatments, PLS-DA revealed a strong separation between both treatments with the fungicide (fungicide and fungicide + bactericide) and the control and bactericide only treatments. Pairwise tests confirmed this result and showed a marginally significant difference between the blends emitted by plants treated with both fungicide treatments and the others (0.05 < P < 0.1). Emission rates for all compounds did not significantly change though there was a trend toward higher emissions with fungicide treatments, and are depicted in Fig. S1 (in supplementary files).

Figure 1.

Partial Least Squared – Discriminant Analysis (PLS-DA) performed on all compounds (ng.g⁻¹.h⁻¹) before (left) and after (right) the different treatments: control (C), bactericide (B), fungicide (F) and bactericide + fungicide (FB). PLS-DA were performed after auto-scaling and log transformation of the dataset. Pairwise tests indicate a difference between treatments containing fungicide (fungicide, fungicide + bactericide) and the others (control and bactericidal) after the application, with n = 5 and 0.05 < p-value < 0.1.

Application of the fungicide resulted in four new compounds appearing in the blend: cyclohexyl cyanide, cyclohexanone and 2 unknown compounds (Table 1; Fig. S2 and S3, in supplementary files). These four compounds were emitted in greater amounts after application of the antifungal treatments than before the application (0.05 < P < 0.1). However, the appearance of those four compounds did not significantly change the total emission rates between treatments (P > 0.05, 22.1 ± 6.9 ng.g_DW⁻¹.h⁻¹, 19.4 ± 8.6 ng.g_DW⁻¹.h⁻¹, 85.2 ± 44.2 ng.g_DW⁻¹.h⁻¹, 124.8 ± 74.8 ng.g_DW⁻¹.h⁻¹ for control, bactericide, fungicide and fungicide + bactericide, respectively, Fig. S1 in supplementary files), though the standard deviation was high for the five replicates and a tendency toward higher emissions for the fungicide treatments was noted, as mentioned above.

Table 1.

Four compounds indicated through VIP scores as separating fungicide treatments from other treatments. Paired Wilcoxon tests according to the time of sampling (before and after application) for fungicide (F) and fungicide+bactericidal (FB) treatments with n = 5.

	Compounds
	Cyclohexanone	Cyclohexyl cyanide	Unknown compound 1	Unknown compound 3
RT (min)	16.289	22.626	35.791	45.453
Kovats index	897	1013	1259	1466
Formula	C6H10O	C7H11N	C10H12O2	C10H13NO
Type of compound	Ketone	N-containing compound	Oxygenated monoterpene	N-containing compound
VIP scores	C1: 2.5; C2: 2.4	C1: 2.5; C2: 2.4	C1: 2.5; C2: 2.4	C1: 2.5; C2: 2.4
P-valueF	0.063	0.063	0.063	0.063
P-valueFB	0.063	0.063	0.063	0.063

Discussion

In our study, four novel compounds were found in the blend of BVOCs emitted from leaves after the application of the fungicide treatment. This observation could be the result of three different mechanisms. Firstly, the results could suggest that, unlike bacteria (see below), fungi may inhibit the emission of certain BVOCs. Secondly the novel compounds observed may have been taken up and metabolized by fungi living in the phylloplane of the B. nigra leaves. Indeed, it has earlier been shown that microorganisms have the ability to consume BVOCs and metabolize them.⁹ Thirdly, it is possible that the fungicide used in this study could have disturbed the metabolism of plants, suggesting possible adverse effects of pesticide treatments on plants in agriculture.

One of the novel compounds observed, cyclohexyl cyanide, could be directly emitted by B. nigra leaves, as it is well documented that Brassicaceae species emit nitriles.¹¹ The function of cyclohexanone is still unclear, but previously it has been shown that the emission of this compound is bound to bacterial activity.¹² It is also plausible that cyclohexyl cyanide has a bacterial origin, since a similar compound, 4-cyanocyclohexene (known as cyclohexenyl cyanide), was found to be emitted by bacteria.¹³ Both of the unknown compounds were difficult to identify based on compound libraries and retention indices. The unknown compound 1 was annotated as β-thujaplicin (mass spectra presented in supplementary files S4). However, the experimental Retention Index (RI, 1229) for this compound did not match with published RIs for β-thujaplicin in literature (1475, https://webbook.nist.gov/chemistry/). The unknown compound 3 was annotated as 1,4-benzoquinone,2,6-diethyl-,1-imine. As for the unknown compound 1, this annotation is not clear and should be confirmed by analytical standard injection.

Our results showed that bactericide treatment did not affect the quality or quantity of emissions. These results differ from those of Peñuelas et al.⁷where they showed a decrease of Sambucus nigra floral emissions after bactericide treatment. In their study, they used three different bactericides at relatively high concentrations that have a broad effect on bacterial communities, whereas in our study we used a lower concentration of a single bactericide. This difference in protocol could be responsible for the different results obtained. Thus, the new compounds detected in the volatile blend after removal of the fungi could have been emitted by the plant, but could also have been emitted by bacteria remaining on the plant if the bactericidal treatment failed. However, it should also be noted that bacterial communities of flowers and leaves are likely to be different and consequently the role of bacteria in floral and leaf emissions could vary substantially.¹⁴

It seems that there is a close relationship between the fungal communities existing on the plant leaf surfaces of B. nigra and the composition of BVOC emissions (through inhibition and/or metabolism). More studies are now required to determine if the new compounds observed are emitted directly from the plant and are linked to microbial metabolism or if other microorganisms emit them directly. Future studies should also focus on the microbes likely to be controlling the BVOC emissions, e.g. by characterizing the microbial communities present with and without the different treatments using modern molecular methods.^15,16 It has also to be noted that this experiment was performed in controlled conditions which could have resulted in impoverished microorganism communities living on the phylloplane compared to natural conditions.¹⁷ Thus, it is possible that different results could have been observed in the field. Performing this experiment under more natural conditions would enable further elucidation of the ecological roles of microorganism-induced changes in VOC blends.

As suggested by Peñuelas and Terradas⁶, microbial communities seem to play an important role in the blend of BVOCs emitted by plants. Therefore, it is crucial to improve our understanding of the relationships between phylloplane communities and plants, especially in terms of BVOC emissions. Changes in those communities, by fumigation with bactericides and fungicides that are largely used in modern agriculture, could lead to modification of plant BVOCs. Consequently, the changes could disturb a plant’s interactions with other organisms such as other plants, herbivorous insects or pollinators. All these hypotheses require further investigation.

Funding Statement

This work was supported by the Academy of Finland through decision number [309425].

Acknowledgments

We thank Jaana Rissanen and Henna Toppinen for their help during measurements and analyses. We also thank Erik Poelman for providing seeds for this study.

Author Contribution

AS, JDB and CB designed the experiments. AS and PM conducted the research, collected and analyzed the data. AS and JDB wrote the manuscript.

Disclosure of Potential Conflicts of Interest

No potential conflicts of interest were disclosed.

Supplementary Material

Supplemented data of this article can be accessed publisher’s website.