Results

In planta protective effect of HPOs against phytopathogens

The capacity of HPOs to induce SR in A. thaliana against B. cinerea was tested under controlled conditions by treating plant roots with HPO solutions and inoculating plant leaves with the pathogen. The disease severity with HPO treatment was measured after 4 days, compared to controls (treatment with water containing 1% DMSO) and we attributed a disease severity scores from 1 to 4 representing no symptoms (score 1) to full development of the disease (score 4). Treatment with both HPOs provided similar obvious disease reduction as shown by the decreased size of the lesions (Figure 1 ). About 40% of the plants had no symptoms which leads to disease severity scores mainly between 1 and 2. Another host-pathogen couple (Solanum tuberosum var. Bintje infected byPhytophthora infestans) was also considered in order to check if the protective action of HPOs is pathosystem-dependent. It was found that 13-HPOT offers more than 50% protection against the pathogen (seeFigure S1 ). Given the experimental design (treatment and infection on two different plant organs or upstream of an infection), this cannot be a direct biocidal effect but rather a systemic signalling in the plant. One can therefore wonder about the signalling mechanism initiated by HPOs and more particularly their initial perception by the plant cells and the responses they induce. HPOs were efficient in both pathosystems studied, suggesting that the mechanism of perception is not dependent on specific recognition receptors.

Perception of HPOs and early defenses responses activation

Very often, ROS production is a biphasic process with a first transient phase within minutes after the infection and a second more intense and sustained phase that can last for many hours (Wang et al. , 2019). This first wave, which is linked to the activation of early defence responses, has been investigated to determine whether the two HPOs are perceived by plants and can induce an immune response.
Cell suspensions cultures are a valuable model system for studying elicitor-induced defence reactions in plants and they easily allow studying early signalling events like oxidative bursts (Khonon et al. , 2011; Jogaiah et al. , 2019). Here, photoautotrophicA. thaliana cell suspensions were used to detect H2O2 production after treatment with HPOs. Extracellular ROS accumulation was detected using a luminol-based assay (van Aubel et al. , 2016; Monnier et al. , 2018). The elicitor FytoSave® was used as positive control as its active substance, COS-OGA made of pectin-derived oligogalacturonides (OGA) and chitooligosaccharides (COS) (Cabrera et al. , 2010), is known to induce a significant production of extracellular ROS at a concentration of 25 ppm (Ledoux et al. , 2014; van Aubel et al. , 2016, 2018). In our experiment, a range of six different concentrations (0.5 µM to 100 µM) was tested and ROS production was monitored for 90 min.
The ROS production after 13-HPOD or 13-HPOT treatment was concentration-dependent (Figure 2A). It was higher than the negative control but also higher than the positive control, for concentrations 50 µM and 100µM. In those cases, the response to 13-HPOT treatment is higher than the 13-HPOD one. Also, the oxidative burst peak occurred quicker (30 min instead of 40 min) for 13-HPOT than for 13-HPOD (Figure 2B). The same conclusion was drawn from experiments performed on foliar disks of A. thaliana , a plant model closer to the reality (see supplementary data – Figure S2 ).
Comparatively to COS-OGA, the kinetics of ROS production induced by HPOs in plant cells was slower (within 5 min only for COS-OGA vs 30-40 min for HPOs) but the oxidative burst lasted longer with a total duration of 60-70 min before returning to the basal level. Such long lasting response profile was also observed with synthetic RL bolaforms, for which it was suggested that their perception occurred via the lipid fraction of the plasma membrane (Bahar et al. , 2016; Luzuriaga-Loaiza et al. , 2018; Come et al. , 2021). On the contrary, the elicitor flagellin, known to be perceived by membrane pattern recognition receptors (Meindl et al. , 2000; Smith and Heese, 2014), induces a quicker oxidative burst initiated within 4-6 min with a peak at ~10min (Yu et al. , 2017), similarly to the OGA in the COS-OGA composition (van Aubel et al. , 2013; Smith and Heese, 2014). Moreover, the induction of ROS production occurs at concentrations much higher than concentrations usually active when a proteic receptor is involved as for flagellin which still binds to its receptor at femtomolar concentrations (Meindl et al. , 2000; Zhao et al. , 2010). The observation of different kinetic profiles and different concentration ranges for HPOs comparatively to classical elicitors would suggest that a proteic receptor is not directly involved in their recognition. Due to the ability of HPOs to interact with PPM lipids (Deleu et al. , 2019), we hypothesized that the HPOs would rather be recognized via the lipid phase of the membrane.

Changes of PPM biophysical properties induced by HPOs

The interaction of HPOs with PPM characteristic lipids has already been found to modify the lateral organization of membrane bilayer in terms of lipid domain size and distribution (Deleu et al. , 2019). It is also known that the plant sphingolipid GluCer is a privileged partner for the interaction and that 13-HPOT has a higher interaction affinity than 13-HPOD.
In the present study, further analysis of the effects of HPOs interaction with lipids on PPM structure was carried out. Simplified biomimetic models with two different lipid compositions were studied, the first mimicking the PPM, namely PLPC:sito:GluCer (60/20/20), and the second made of d62DPPC, a classic deuterated model.

Effect of HPOs on membrane permeability

First, we have investigated the ability of HPOs to permeabilize model membranes by measuring the release of calcein. If the membrane is permeabilized by a bioactive molecule, the self-quenched calcein initially encapsulated within the LUV is released into the external medium and gives rise to an increase of fluorescence emission. Very little permeabilization effect was observed (values less than 10%) for both HPOs on the PLPC:sito:GluCer membrane model (Figure S3 ) suggesting that HPOs would not derive their mode of action from a mechanism of solubilization of the membrane or pore formation, but rather a more subtle modification of the membrane organization that could lead to the activation of a signalling cascade.

Change in bilayer fluidity induced by HPOs

The effect of HPOs on the bilayer fluidity was investigated by monitoring the lipid phase-dependent emission spectrum shift of Laurdan, a fluorescent probe that readily locates at the hydrophilic/hydrophobic interface of bilayers (Harriset al. , 2002; Sanchez et al. , 2007). Its fluorescence depends on the physical state of the environment. When present in a gel phase bilayer, its maximum fluorescence intensity is close to 440 nm emission wavelength. When the bilayer is in a fluid state, the Laurdan maximum fluorescence is observed at higher wavelengths (around 490 nm). This ”red-shift” phenomenon is due to a higher quantity and mobility of the water molecules located around the probe. This is directly related to the lower order within the bilayer and is measured by the Generalized Polarization (GP): a decreasing GP value corresponds to a higher fluidity of the bilayer (Parasassi et al. , 1991). This method has been previously applied for investigating the effect of drugs, natural herbicides or other elicitors on lipid membrane organization (Deleuet al. , 2013; Sautrey et al. , 2014; Lebecque et al. , 2019; Furlan et al. , 2020).
The effect of HPOs on PLPC:sito:GluCer MLV membrane fluidity was investigated for a range of temperatures from 20°C to 50°C (Figure 3 ). In presence of 13-HPOT, the Laurdan GP values decreased significantly compared to those observed on pure MLVs with no significant effect of temperature. This indicates a fluidifying effect of 13-HPOT on the bilayer. On the contrary, 13-HPOD did not induce significant change in lipid order at any temperature as its curve almost superimposes to that of pure PLPC:sito:GluCer vesicles.

Effect of HPOs on the bilayer transversal structure

The effect of HPOs on the transversal structure of PPM was analysed by NR, a technique of choice to study the transverse structure of layered samples within a few Å resolution (Mattauch et al. , 2018) and to evidence the structural effects of the interaction of incoming molecules on biological membranes (Rondelli et al. , 2016, 2018). Neutrons interaction with matter depends on the isotopic species. Therefore, neutron-based experiments can profit by the use of deuterated molecules to enhance the visibility of molecules within a mixed complex system. As the lipids representative of the PPM are not commercially available in their deuterated form, d62DPPC was used to form SLB and to highlight the presence and location of the H-bringing HPOs within the membrane.Figure 4A-D show the reflectivity curves together with their fittings in two contrasts and the corresponding fit parameters are summarized in Figure 4E . NR spectra were not drastically changed by the addition of HPOs. However, the data analysis revealed that HPOs always insert into the outer hydrophilic leaflet of the d62DPPC SLB without flipping into the inner layer attached to the silicon block (this is evident from the variation of the external polar leaflet SLD and roughness, as reported in Figure 4E). A slight modification of the SLD profiles was observed while adding 13-HPOT but not with 13-HPOD. This gave rise to a small but significant decrease of the membrane thickness (approximately 2 Å) and roughness without modification of the solvent penetration and of lipid chains SLD,i.e ., no alteration of the bilayer nor external molecules deep penetration. To confront these results, obtained on a d62DPPC bilayer, to a more realistic PPM model, another experiment was performed with the deuterated version of 13-HPOD and the ternary mixture of non-deuterated lipids representative of PPM. It confirms that 13-HPOD interacts with PPM SLB and localizes on top of the outer leaflet without no major change of the membrane organization as observed from SLD profile and NR spectra (Figure S4 andTable S2 ).

Lateral erosion of plant lipid bilayers by HPOs

To further analyse the effect of HPOs on the lipid bilayer organization, atomic force microscopy (AFM) was used to investigate their impact on the lateral nanoscale morphology of SLB. As shown in Figure 5A , the ternary mixture of plant lipids reconstituted in SLB did not reveal any phase separation within the thickness resolution limit of AFM (0.1 nm), rather homogeneous smooth patches (bright areas) distributed on the mica (dark areas). Though large patches were mainly found to cover the entire scanned area, defects in the SLB were used as a “visualization control” to confirm the presence of the lipid bilayer. Its thickness of ~4-5 nm, determined by measuring section profiles on the AFM height images, is in agreement with previous studies (Figure 5B ) (Dufrêne and Lee, 2000; Mingeot-Leclercq et al. , 2008). The presence of the lipid bilayer was further confirmed by recording AFM force curves on areas of high vs low heights. A typical breakthrough of the lipid bilayer in the bright areas was observed while no such force signature was found in dark areas without lipid bilayers and associated with mica (Figure 5C ).
After confirming the presence of the SLB, the sample was incubated with either 13-HPOT or 13-HPOD and AFM images were recorded every 10-15 min on a defined area. Incubation 13-HPOT or 13-HPOD resulted in a time-dependent alteration of the lipid patches (Figure 5D ). Results showed that very small SLB patches (green arrows) disappeared after the addition of 13-HPOT, but that it had not a drastic impact on the large ones. On the contrary, 13-HPOD completely removed part of a large angular domain after 75 min (see green arrows). Nonetheless, after 75 min treatment with 13-HPOT, most of the lipid domains were thinner by approximately 2 nm as compared to the initial ones, suggesting that 13-HPOT flattened lipids or part of the upper leaflet in a time- and zone-dependent way, which was not observed for 13-HPOD (Figure 5E-F ).
In brief, AFM studies revealed three major effects of HPOs on plant mimetic lipid bilayers (i) ”erosion” of angular protrusions of large lipid domains, (ii) total erosion of small domains, and (iii) reduction in the thickness of the bilayer between 0.5 and 2 nm. 13-HPOT has also a greater effect on membrane organization and bilayer thickness than 13-HPOD.