Xylella fastidiosa subsp. pauca and olive produced lipids moderate the switch adhesive versus non-adhesive state and viceversa.

Global trade and climate change are re-shaping the distribution map of pandemic pathogens. One major emerging concern is Xylella fastidiosa, a tropical bacterium recently introduced into Europe from America. In last decades, X. fastidiosa was detected in several European countries. X. fastidiosa is an insect vector-transmitted bacterial plant pathogen associated with severe diseases in a wide range of hosts. X. fastidiosa through a tight coordination of the adherent biofilm and the planktonic states, invades the host systemically. The planktonic phase is correlated to low cell density and vessel colonization. Increase in cell density triggers a quorum sensing system based on mixture of cis 2-enoic fatty acids—diffusible signalling factors (DSF) that promote stickiness and biofilm. The lipidome profile of Olea europaea L. (cv. Ogliarola salentina) samples, collected in groves located in infected zones and uninfected zones was performed. The untargeted analysis of the lipid profiles of Olive Quick Decline Syndrome (OQDS) positive (+) and negative (-) plants showed a clustering of OQDS+ plants apart from OQDS-. The targeted lipids profile of plants OQDS+ and OQDS- identified a shortlist of 10 lipids that increase their amount in OQDS+ and X. fastidiosa positive olive trees. These lipid entities, provided to X. fastidiosa subsp. pauca pure culture, impact on the dual phase, e.g. planktonic ↔ biofilm. This study provides novel insights on OQDS lipid hallmarks and on molecules that might modulate biofilm phase in X. fastidiosa subsp. pauca.


Introduction
Xylella fastidiosa (Xf) is one of the top 10 plant pathogenic bacteria [1] and is the cause of an environmental emergency within the European Union (EU). Xf has infected a broad host range [2] of plant species; symptoms vary depending on the combination of the host plant and Xf strain [3]. At present, X. fastidiosa subsp. pauca, associated to the OQDS, and three other subspecies, fastidiosa, multiplex, and sandyi have been identified in Europe [4][5][6][7]. Xf subsp. pauca was initially introduced from America into Southern Italy and recovered in olive trees [8,9]. This pathogen is associated to the OQDS which has caused losses up to €390 M of the national olive oil production, in the last three years in Italy [10]. Xf is an obligated vector-transmitted pathogen and a xylem-limited bacterium. Inter-host transmission is mediated by xylem sap-feeding insects. The biology of invasion has been described in the grapevine [11]. From the point of entry in grapevine, Xf moves along the xylem, attaches to its walls and, through a tight coordination of the adherent biofilm and the planktonic states, invades the host systemically [12,13]. The planktonic phase is correlated to low cell density and vessel colonization [14]. The colonization during the exploratory phase (planktonic) is correlated to low cell density and symptoms progression in grapevine. Increase in cell density triggers a quorum sensing system based on cis 2-enoic fatty acids-DSF that promote stickiness and biofilm formation [15]. Xylem vessels are occluded by the combined effect of bacterial biofilm and plant defences such as tyloses and formation of amorphous gels coating xylem vessel walls, causing characteristic symptoms such as leaf scorch [11,16]. DSF can modulate gene-expression in Xf [17,18]. DSF-mediated quorum sensing determines a) degradation of the pit membranes to enable cross-vessel diffusion; b) twitching motility of Xf cells; c) adhesion to the xylem surface and biofilm formation. The earlier stage of infection in Pierce disease, consists of evasion by the pathogen of the plant innate immune response and colonization of its vessels, de facto limiting opportunities for reacquisition by the feeding insect vectors [19]. At this stage, Xf is not detected by the plant as a biotic stress, but rather as an abiotic stress (drought and dehydration) [20]. At later stages, the biofilm-based phenotype, consisting of a high density of Xf cells, facilitates reacquisition by the vector and dissemination into other hosts. Only at this point, the plant recognizes the pathogen and mounts an immune response ineffective at preventing Xf colonization and symptoms [21,22]. DSF act as coordinators of this dual activity of Xf, allowing the switch from the early stage (planktonic endophytic lifestyle) into the later one (sessile insect-acquisition stage) [11,23]. Lipids appear to be central in this pathogen interplay between the two phenotypes. In Pseudomonas aeruginosa, other lipids, namely the oxylipins, act as hormones for controlling the switch among the different stages of bacterial lifestyle: planktonic, twitching, and biofilm. In P. aeruginosa, the oleic acid-derived oxylipins control the virulence in the host and function as autoinducers of a novel quorum sensing system mediating cell-to-cell communication in bacteria [24,25].
To provide a broader view of the lipids emerging in OQDS symptomatic trees infected by Xf, we analyzed the lipidomic profile of samples of 60-y.o. Olea europaea L. cv. Ogliarola salentina symptomatic and symptomless for the OQDS. The profile of 437 lipid compounds was assayed: 186 were found to be differentially accumulated in OQDS positive individuals of which 90 were further characterized and quantified by MS/MS spectrometry. Among these, we identified ten compounds, more abundant in Xf+ and OQDS+, that are novel hallmarks of Xf symptoms and infection of olive trees. Importantly, we found that these lipids modulate Xf lifestyle under in vitro conditions.

Study site and sampling procedures
Sampling was carried out in the Apulia region, in October 2017, on 120 individuals of Olea europaea L. cv. Ogliarola salentina (60 years old), 60 showing OQDS symptoms (+) and 60 OQDS symptomless (-). The OQDS+ individuals were collected in an olive grove in Copertino in the infected area of the Lecce province (40˚16'5.56" N 18˚03'15.48" E); the OQDS-individuals were sampled in Grottaglie, Taranto province (40˚32'12.98" N 17˚26'14.03" E) in an area regarded by the phytosanitary service of the Apulia region as still unaffected by the pathogen. Trees were identified as symptomatic or symptomless following the criteria previously reported [26]. Mature shoots from each OQDS+ tree was collected from sections of the canopy on branches that showed desiccation and dieback. The samples consisted of "1-or 2-year-old twigs (ca.0.5 cm in diameter) from which cuttings of 15-20 cm were prepared from the portions close but yet unaffected by the withering and desiccation phenomena" [26]. Xylem tissue was then recovered, after removing the bark, and processed. The same procedure was performed for each OQDS-tree. The obtained 120 samples were separately lyophilized. 12 pools of equal weight (1g) were generated from the lyophilized samples, namely:6 pools (each representing 10 individuals) from OQDS+ and 6 from OQDS-. For a comprehensive lipidomic analysis, the pools included the overall analytical complexity of our samples [27,28]. OQDS + and OQDS-samples were molecularly assayed via real-time PCR in technical triplicates [3,29] to verify the presence of Xf and thereon defined as Xf+ and Xf-samples (S1 Fig).

Lipids analysis
Xylem tissue (1,0 gr) was recovered and lipids extraction and analysis were performed as previously reported [30]. Xf+ and Xf-samples were assayed with the internal reference standards tricosanoic acid, glyceryl tripalmitate d31, and 9-HODEd4. The analysis was carried out at a final concentration of 2μM. The samples were analysed by untargeted lipid analysis conducted with a G6220A TOF-MS, (Agilent Technologies, USA) operating in negative and positive ion scan mode as previously described [30]. A sub-group of lipid classes was analysed (fragmentation analysis) by LC-MS/MS (Triple Quadrupole; 6420 Agilent Technologies, USA) as reported [30]; multiple reaction monitoring (MRM) methods were adopted to analyse the most abundant lipid entities (S1 Table). MRM data were processed using the Mass Hunter Quantitative software (B.07.00 version, Agilent Technologies, USA). The mass spectrometry analyses were performed three times, each time in technical triplicate (n = 9). PCA and significance-fold change analysis (Volcano plot) for untargeted LC-TOF/MS results were performed trough Agilent Mass Analyzer software. Significance tests (T-Student Test, p<0.05) and plots of MRM and SIM results were performed trough R software.

Quantification of biofilm formation
An in vitro test was made to assess the effect of free fatty acids, diacylglycerides, and oxylipins on biofilm formation of Xylella fastidiosa subsp. pauca strain De Donno (CFBP 8402). Xf subsp. pauca biofilm formation was evaluated as previously described [31] with some modifications. Briefly, a pure culture of the bacteria was grown for 7 days on PD2 plates, scraped and resuspended in PBS. 10μL of cell suspension (A600 = 0.5 OD) was inoculated in a sterile glass tube containing 1mL of PD2. The free fatty acids (FFA) (Sigma-Aldrich, USA) or diacylglycerides (1,3-Dilinoleoyl-rac-glycerol; 1,3-Diolein) (Sigma-Aldrich, USA) or oxylipins [Cayman chemicals, USA or 7,10 DiHOME and the mix (7,10 DiHOME; 10-HOME) kindly provided by Dr. Eriel Martínez and Javier Campos-Gómez (Southern Research Center, AL, USA] were added to the medium at desired concentrations when required as reported [24]. After 11 days of incubation (28˚C; 100 rpm), the total number of cells-planktonic growth (cells in suspension) and biofilm growth (cells adhered to the substrate) was estimated. The lipid compounds were divided into two groups: those dissolved in EtOH (with EtOH) and those dissolved in water (no EtOH). Spectrophotometric absorption of Xf subsp. pauca cultures was used to measure growth (600 nm) and biofilm formation (595 nm). For compounds dissolved in EtOH, references with a corresponding concentration of EtOH but without the compound were used as background. Impact on growth / impact on biofilm are defined as the amount of growth/ biofilm formation minus the background, normalized to the growth/biofilm formation in a medium in the absence of added EtOH or lipids. A positive impact indicates values more than no-lipid controls (i.e. improved growth or biofilming), whereas a negative impact indicates values less than no-lipid control (i.e. inhibited growth or biofilming). EtOH at similar molarity to lipids was additionally used as a point of reference of how Xf subsp. pauca growth and biofilms were affected. The experiments were performed in biological triplicate for each treatment and carried out three times (total repetitions per treatment n = 9). Multiple comparison with Kruskal-Wallis test (p value< 0.05) and Fisher's LSD post-hoc test, with Bonferroni correction, were run on R software to individuate significant (p<0.05) groupings within the different treatments.  (Fig 1). PCA highlighted two primary components (X-axis: 60.48%; Y-axis: 12.05%) that separated the Xf+ and Xf-clusters (Fig 2). Volcano plot analysis on positive-ion scans provided a few significantly modified compounds (S2 Fig).

Discussion
In the host plant, Xf colonizes the xylem vessels and elicits their occlusion [11,23,32]. Xf maintains a lifestyle that switches bidirectionally from adhesive to non-adhesive cells phenotype and vice versa [14]. Lipids are crucial signals that can modulate the pathogen virulence and play important roles during the infection process [15,33]. Among lipids, oxylipins can modulate biofilm formation and virulence in the Gram-negative pathogen P. aeruginosa [25]. Most information on the biology of Xf is mostly referred to one of its subspecies, i.e. fastidiosa, in the context of Pierce's disease and, regarding lipids involvement, to DSF [21]. Recently, we showed that during the infection of the model organism Nicotiana tabacum by Xf subsp. pauca, several lipids were differently accumulated in infected versus healthy plants; DSF-like compounds apart, oxylipins emerged as hallmarks of pathogenic invasion of host tissues [30]. In olive trees, reports are available on the molecular basis of Xf subsp. pauca invasion [26,32]. More recently, Cardinale and colleagues [34] demonstrated that Xf subsp. pauca forms biofilm in xylem vessels of OQDS+ trees (cultivar "Ogliarola di Lecce"), pinpointing the role of bacterial aggregates in vessel occlusions in naturally infected olive trees. Notwithstanding the importance of lipids in Xf lifestyle, no report employs a lipidomic approach to differentiate OQDS+ from OQDS-trees. We explored for the first time the lipidomic profile of olive trees naturally infected by Xylella fastidiosa. Lipids play a key role in plant disease. In this work, we check by untargeted analysis all the lipids that change under the pathogen pressure; moreover, PCA of the lipid entities allowed to differentiate the infected trees from the uninfected ones. Since the OQDSsamples clustered differently from OQDS+ samples, we suggest that the differential formation of lipid entities can be modulated by the presence of Xf producing OQDS symptoms. Lipids are essential constituents of the cells involved in different biological functions; this study could pave the way for developing a diagnostic tool targeted on the lipids that differentiate the infected plants by non-infected ones. In this regard, the Volcano plot analysis of the extracted entities from olive trees affected by OQDS, highlighted that 186 lipids compounds were specifically formed in symptomatic samples. More intriguingly, 10 compounds emerged as statistically able to discriminate Xf+ from Xf-olive trees. Considering that this part of the study gathers data straight from natural infection processes, we suggest these specific lipids as hallmarks of Xf infection of olive trees and propose to extend this type of approach to other hosts affected by this fastidious bacterial pathogen.
In olive trees, OQDS symptoms are closely related to xylem vessel occlusions caused by biofilm formation from Xf [34]. Lipid metabolism has a strong impact on plant-host interaction and, despite the proceedings in plant pathology, many important questions remain unanswered about the functional diversity of lipids and the mechanisms bacterial cells employ to coordinate their response to lipids modulation. In this study, we investigated how lipids differently accumulated in OQDS+ samples affected the lifestyle of X. fastidiosa subsp. pauca. To identify their role in Xf biofilming [18,25], we tested in vitro the oleic acid-derived 7,10-diHOME alone as well as in conjunction with 10-HOME (DOX-oxylipins); the free fatty acids C18:1 and C18:2 and their diacylglycerols (DAG 36:2 and 36:4); the LOX-derived 9-oxylipins (9HODE, 9-OXODE, 9-HOTRE) and 13-oxylipins (13-HODE, 13OXODE and 13-HOTRE).
Our working hypothesis on the effects of these compounds on Xf biofilming stemmed from the available information on their biological role and was inspired by Roper and colleagues' intuition: "the bacteria induce an autoimmune-like syndrome" [14]. More specifically, the oleic acid is among the modulators of quorum sensing in Xf [15]; in other bacterial pathogens its ratio with the linoleic acid could promote virulence [33], whilst DAG associated compounds (e.g. DAG 36:2) lead to establish an appropriate defence response by inducing defence-signalling molecules [33]. As reported elsewhere, oleic acid-derived DOX-oxylipins are instrumental in moderating the lifestyle of another opportunistic pathogen: P. aeruginosa [24]. Concerning LOX-derived products, in animal systems, 9-oxylipins (e.g. 9-HODE) may display a pro-inflammatory effect, whereas 13-oxylipins (e.g. 13-HODE) has an anti-inflammatory one [35]. Some authors suggested using 13-to 9-oxylipins ratio as a marker of the inflammatory status in murine models [36]. 9-and 13-oxylipins play a similar, apparently antagonistic role in fungal-plant interaction [37]. Our results show that 18:1 DOX-oxylipins (e.g. 7,10-diHOME) hold back biofilm formation, while LOX-derived 9-oxylipins (e.g. 9-HODE) stimulate biofilm formation. Regarding a possible antagonistic role for 9-and 13-oxylipins, we noticed a similar antagonistic effect in biofilming (i.e. 9-and 13-HOTRE).
Our hypothesis provides the following scenario: plants produce oxylipins in response to several stresses, including which pathogenic insult [38]. Oxylipins, other than functioning as plant defence signals [39], may display an effect toward pathogens in modulating their lifestyle, or a biocidal one [40,41]. Notably, LOX-derived products could interact with the membranes of pathogens altering their organization and disturbing cell growth [42]. Further studies are needed to clarify how Xf can adapt to oxylipins-modulated plant stress response and exploit it to trigger its biofilming.
We highlight for the first time the different accumulation of FFA, oxylipins and DAG in OQDS+ and Xf+ olive trees and we suggest their role in the modulation of Xf subsp. pauca biofilming. We encourage researchers to investigate oxylipins as new targets for the development of treatments for OQDS in line with the Roper team suggestion: "design therapeutics that target the dispersal state or encourage adhesion to lock the bacterial population into its self-limited state permanently" [14].
A more straightforward outcome of our results leads to exploit lipids as markers for developing diagnostic methods based on both destructive strategies, such as molecular tests (lipids related genes) and/or MS spectrometry analysis, and on non-destructive tests such as volatilome of Volatile Organic Compounds (VOCs) derived lipids, as recently suggested [43,44].  Table. Compound identification from LC-TOF untargeted analysis. Each row provides the compound putative identification and values explaining differences between Xf-positive and Xf-negative samples, i.e.: p-value, corrected p-value, Fold Change of Xf-positive versus Xfnegative, abundance difference of Xf-positive versus Xf-negative in raw and log2 form. Compounds highlighted in yellow were fragmented and characterized through product ion experiments. (DOCX) S3 Table. Normalized peak areas of SIM and MRM analyzed lipid entities. Compounds' peak areas were divided by their internal standard's peak area and then by the maximum value of internal standard peak area. For each compound the fold-change of Xf+ samples versus Xfsamples was calculated, as such the p-value from Student T-test. Compounds with a p-value <0.05 are represented on the upper part of the volcano plot, above the red line corresponding to -log 10 (0.05). (DOCX)