Biocidal Activity of Low Temperature Plasma to Xylella Fastidiosa

The quarantine bacterium Xylella fastidiosa was rst detected in Salento (Apulia, Italy) in 2013 and caused severe symptoms in olives, leading to plant death. The disease, named Olive Quick Decline Syndrome (OQDS), is caused by the strain "De Donno" ST53 of the subspecies pauca of this bacterium (XfDD), which is spread by Philaenus spumarius. The epidemic poses a serious threat to the agricultural economy and the landscape, as X. fastidiosa infects several plant species and there is yet no recognized solution. Research on OQDS is focused on nding strategies to control its spread or mitigate its symptoms. In this context, we investigated the feasibility of using low-temperature plasma and plasma-activated water to kill bacterial cells. Experiments were conducted in vitro to test the biocidal effect of a Surface Dielectric Barrier Discharge plasma on bacteria. The results showed a high decontamination rate even for cells of XfDD embedded in biolms grown on solid media. Application to trees requires protocols and tools that can reach the bacterium in the xylem vessels. Plasma Activated Water was tested as a biocidal agent that can move freely in the xylem network. Results in the liquid culture medium showed complete inactivation of XfDD cells and paved the way to test the strategy on infected plants.


Introduction
Xylella fastidiosa is a xylem-limited gamma proteobacterium whose pathogenic mechanism is based on the occlusion of the xylem vessels. The bacterium lives in bio lm communities embedded in an exopolysaccharide matrix from which clumps of cells are released to colonize infected plants systemically. Infection of new plants is mediated by insect vectors that feed in the xylem, ingest the bacterium and transfer it to new plants during successive feeding activities 1,2 .
The "De Donno" strain of Xylella fastidiosa subspecies pauca haplotype ST53 was discovered in late 2013 in the Salento Peninsula (Apulia region) in olive trees that showed leaf burn, extensive desiccation of branches and plant death 3 . Soon after, this bacterial strain was found to be the causative agent of these symptoms, which were grouped under the name Olive Quick Decline Syndrome (OQDS) 3 . Although this strain infects the two main autochthonous olive varieties, Ogliarola salentina and Cellina di Nardò, it was also found in 595 other plant species, in several cases without showing symptoms 4 .
The discovery of X. fastidiosa in this new environment and the infection mainly of olives has stimulated several research programs that have led to the description of the epidemiology of the infections in the epidemic area Salento Peninsula, the identi cation of Philaenus spumarius as an insect vector, the genomics of the bacterium and the response of olives to the infections 5 .
The X. fastidiosa epidemic in southern Italy required the implementation of a series of phytosanitary measures enforced by European Authorities. These included intensive surveys to delineate the infected area and exclusion of the pathogen from non-infected areas in the Italian peninsula as well as in the other Member States. The results of the surveillance program have con rmed that the infections are slowly spreading northwards into the southern part of Apulia. However, in addition to Italy, where a subspecies multiplex strain of X. fastidiosa has been found in Tuscany 6 , other strains have been detected in France 7 , Spain 8, 9 , Portugal 10 and recently in Israel 11 , highlighting the serious threat that this quarantine bacterium poses to Mediterranean agriculture.
Bacterial cells are generally inactivated by physical and/or chemical sterilization. Physical methods such as heat and gamma radiation and chemical methods such as ethylene oxide are not suitable for curing Xylella-infected plants. Heat can cause irreversible damage to the plant, ethylene oxide is highly ammable and toxic, radiation processes require an isolated location and operator safety, and can cause DNA damage. These drawbacks are driving research for a novel, highly e cient sterilization process.
Surface treatments using this dry plasma technology offer an environmentally friendly alternative to conventional wet chemical methods for killing microorganisms. Extensive multidisciplinary research has been carried out over the last two decades, proving the validity of the application of plasma technology in the broad eld of biology and medicine and antimicrobial clinical treatments against various pathogens.
Low-Temperature Plasma (LTP) offers numerous potential advantages over conventional methods, such as nontoxicity, low operating costs, short treatment time at low temperatures, a signi cant reduction in water consumption during the disinfection process, and applicability to a wide range of commodities 16,18−21 .
In general, the effectiveness of LTP depends on the equipment design and system operating parameters, such as gas composition, ow rate, humidity, temperature, voltage, and frequency [22][23][24] . LTP in ambient air is an excellent source of electrons, positive and negative ions, free radicals, stable transformation products, excited atoms and molecules, and ultraviolet radiation (UV) with antimicrobial activity 25 [29][30][31][32][33][34][35][36][37][38][39] . Hydroxyl radicals (OH) have a direct effect on the cell membranes of microorganisms, which are composed of a bilayer of glycerophospholipids and proteins and are susceptible to their attacks 40,41 . In addition to direct decontamination processes, plasma can also activate stress signalling cascades that enable the self-defence mechanism, especially in plants. RONS produced in response to abiotic and biotic stresses act as signalling molecules 42 and induce the production of secondary metabolites that act as precursors of defence hormones and lead to the activation of defence genes [43][44][45][46][47][48] and the production of defence compounds 49,50  In the present manuscript, LTP and PAW were used to evaluate their antimicrobial activities against the strain "De Donno" of Xylella fastidiosa subspecies pauca haplotype ST53 (XfDD). Due to the di culties of performing experiments in an unenclosed laboratory, we developed a portable device to generate plasma through a Surface Dielectric Barrier Discharge (SDBD) using ambient air with a measured relative humidity of approximately 40%. All treatments were performed in a laminar ow hood. We rst investigate the feasibility of direct decontamination of bacteria by plasma by exposing the cell culture to the spatial afterglow of the SDBD for in vitro culture. The second step was to evaluate the effect of PAW in an XfDD suspension in a water-based solution. The basic idea of the proposed experiments is to nd a possible strategy to achieve the decontamination of the surfaces exposed to the bacteria, such as tools or leaves, by using an atmospheric pressure plasma source and to apply the PAW as a novel tool to control the infection in plants, either by direct exposure of the pathogen to the RONS present in the PAW or by activating the self-defence mechanism of the plant.

Materials And Methods
Dielectric Barrier Discharge The discharge system includes an SDBD reactor, gas feeding unit, discharge energization system, electrical and optical diagnostics, and sample holder suitable for inserting biological samples at a selected distance from the discharge surface. The SDBD reactor shown in panel (A) of Fig. 11, consists of a planar SDBD electrode system placed in a PVC chamber equipped with gas feed input/output ports and a high voltage (HV) interface.
The discharge was built starting from a glass petri dish, in which the electrodes were included (Fig. 11). The Ground electrode was equipped with an aluminium air-cooled dissipator to keep its temperature as close as possible to room temperature. The nickel electrode exposed to the discharge consists of 22 parallel stripes (0. to measure the plasma current, and a voltage probe (100:1@10 MW, bandwidth 1 GHz) to measure the potential drop on a transferred charge measuring capacitor (C = 0.02 µF) inserted between the induction electrode and ground. The reactor was fed with humid air at a xed ow rate of 7 slm. This reactor chamber allows the insertion of 50 mm in diameter standard Petri dishes ( Fig. 11) next to the SDBD surface, with the top border approximately 3 mm far from the HV electrode. The same discharge setup was used to produce plasma-activated water (PAW).
The typical appearance of the HV electrode during the on-time is shown in Fig. 11B. Since we are dealing with quarantine bacteria the discharge was operated under a laminar ow hood (model Aura HZ48), in the authorized lab of the CNR IPSP in Bari (Fig. 11C).

Emission spectroscopy
To perform spectroscopic measurement a slightly modi ed version of the reactor was used to place UV grade quartz frontal windows allowing direct visual control of the discharge area and optical emission diagnostics. Optical emission spectroscopy was performed by collecting light from the SDBD plasma surface through a UV fused silica single-lens bre optic bundle model LG-455-020-3 (3 meters, 190 to 1100 nm, with 19 200 µm bres, 10 mm ferrule at slit end) equipped with UV grade quartz lenses system on the entrance slit of a monochromator. The imaged area collected light is limited by the entrance slit dimension and can reach a maximum dimension of 3 (l) × 5 (h) mm 2 . The light is spectrally resolved by a 300/600/1200 grooves mm − 1 , blazed at 300 nm, and covering the ranges 200-1400 nm. The spectrum is acquired with a Princeton Instruments PI-MAX4 1024i CCD camera equipped with a 1024x1024 pixel sensor (size 12.8 um, active area 13.1 × 13.1 mm 2 ). One CCD image of a spectrum covers a range of approximately 144/65/30 nm respectively for the three different gratings. The intensities of the emission spectra acquired by the ICCD detector were spectrally and intensities were corrected utilizing Halogen (oriel) and calibration lamps (oriel). The spectra were used to evaluate rotational temperatures and vibrational distributions. Spectra were simulated to infer rotational temperature information using the freely available spectroscopic tool Massive OES 65-67 .

PAW characterization
Spectrophotometric measurements Measures were repeated 24 h after the activation of water, in samples treated and kept at 4°C, to evaluate changes in the content during the time.

Electrical Impedance (EIS) and Broadband Dielectric Spectroscopy (BDS)
We performed EIS measurements on untreated deionized water (DIW) and plasma-activated water (PAW) at two different plasma treatment times (5 min and 15 min) using a two-probe cell built on 4 ml glass vials. The electrical feedthroughs of the cell consisted of two glass capillary tubes inserted and sealed through a hole in the top of the vial stopper. The capillary tubes allowed the insertion of two gold plated needles, spaced L = 2 mm apart, into the interior of the vial to make electrical contact with the internal water. The needles were then connected to the measurement apparatus via a BNC cable. Impedance (Zvsf) spectra were recorded using a NOVOCONTROL Impedance Analyzer. The AC voltage signal (V AC ) was xed at 30 mV and the frequency range was between 0.1Hz and 1.0 MHz. It took about three minutes to record one spectrum.
Since the magnitude of the impedance depends on the volume and density of the test sample, we perform the measurements with the same amount (3ml) of plasma-activated (PAW) and untreated deionized water (DIW) to obtain the same comparison. We also repeat the measurements on 15 min treated PAW samples after 24 hours to assess the ageing effect due to the instability of the ion concentration.
The instrument software provided the spectra of the real (ReZ) and imaginary (-ImZ) parts of the impedance and the corresponding phase φ, where φ = arctan (ImZ/ReZ)). We plot the impedance outputs using the Nyquist plots (NPs) i.e. -ImZ vs. ReZ. We then t the impedance of a suitable electrical circuit using the free downloadable software EIS Analyzer 80 . Generally, the NP region on the right (left)side corresponds to the low (high) frequency response. In general, the circuit consists of a suitable combination of resistor R, capacitance C, Constant Phase Elements, Z CPE and a Warburg Element, Z W .
The Z CPE is a non-intuitive impedance expressed by 81 : 1 where Q o −1 is the impedance magnitude modulated by the n value and it is introduced to simulate a nonperfect capacitive response In (1) the Q o −1 term is expressed in Ω 1−n F n with 0 < n < 1; therefore, when n is close to 0 the impedance behaves like a resistor whereas when is close to 1 it behaves like a pure capacitance. The Warburg element Z W is the equivalent circuit element assigned to the diffusion process and it is expressed as: 2 i.e. via a constant phase element where the n value is 0.5. In the complex plane, the diffusion element is recognized as a straight line in the NP with a slope of 45° to the real axis.
To gain better evidence on the ionic concentration and conductivity useful for the understanding of the PAW redox activity level, we represent the impedance data via the dielectric loss factor tanδ expressed via the ratio between the real ε' and imaginary ε" part of permittivity i.e. 82 : The dielectric loss relates directly to the AC conductivity and then to the PAW pro-oxidant environment feature since: The expr. 5 shows that the AC conductivity is composed of a frequency-independent part associated with the ionic DC conductivity and a component that depends on the frequency of the voltage signal. In general, the σ vs ω is characterized by the region attributed to electrode polarization effects in the lowfrequency range followed by a plateau in the mid-frequency range (i.e., after the tanδ peak where the σ not depends on the signal frequency and is attributed to the ionic DC conductivity σ DC and then by a region attributed to ionic hopping, which generally follows Johnsher's power law with 0 < n < 1 83 .
Following the Macdonald-Truckam model 84 the peak value of the loss factor and the corresponding angular frequency relates to ions diffusivity D: 6 Since σ DC = n q µ ( n the ionic concentration, q the elemental charge and µ the mobility of the ions) and the diffusivity D = k B T/qµ with k B the Boltzman constant and T the temperature, it follows that the ionic concentration can be determined via: Xylella fastidiosa growth conditions and plasma treatments X. fastidiosa subsp. pauca strain ''De Donno'' ST53 was used in this study 85 . Bacteria were cultured on BCYE agar plates for 15-20 days at 28°C.
Plasma treatments were carried out on bacteria that directly grow on solid (agar amended) media or after suspension in PAW.
Treatment on agar surface-grown bacteria XfDD was currently grown on BCYE agar plates at 28°C and the inoculum was regularly renewed every 7 days on new plates. Cells were collected from the agar surface with a sterile loop, resuspended in 1x PBS buffer and adjusted, before plating, to a nal concentration of 4x10 8 CFU ml − 1 (OD600 0.5).
Trials were carried out using different bacteria concentrations, from 10 7 to 10 3 CFU ml − 1 after resuspending cells scraped from an agar plate in PBS. Forty microliters of bacteria were plated by spreading on the agar surface using a sterile loop. Preliminary tests were carried out to nd the best conditions of treatments, particularly referring to the initial bacteria concentration, the pre-culture of the bacteria before plasma exposure and duration of the exposure. To this aim, bacteria were pre-cultured for 1 and 5 days before being exposed to LTP which was discharged for 10, 100 and 200 s. LTP untreated cells grown in the same conditions were used as control. The plates were incubated at 28°C for up to 30 days and the antibacterial activity was assessed using the viable plate count method 86 . The number of CFU ml − 1 in the control plate was determined by multiplying the number of colonies on a dilution plate by the corresponding dilution factor. Only plates (or replicate plates from the same dilution) with 30-300 colonies were counted 87 .
In addition, the cumulative effect of the LTP was evaluated by applying the SDBD plasma a) one time a day for 1 week, and b) one time a week for 3 weeks at the maximum exposure time of 200 s. In these trials XfDD was pre-cultured for 1 and 7 days in a) and b) respectively, before being exposed to LTP treatment.
Plates were preliminarily treated with LTP for an exposure time of 200 s and successively inoculated with XfDD to evaluate if the plasma treatment could alter the chemical/physical properties of the BCYE agar substrate and therefore inhibit the growth of the bacterium.
Each assay was performed in triplicate, and each experiment was repeated at least two times.

Treatment cells with Plasma Activated Water (PAW)
Deionized water (DIW) (18.2 MΩ·cm at STP) was obtained from a Millipore Direct-Q 3UV system. 500 ml of DIW were stored in a container in an ambient atmosphere for one day. This procedure enabled the concentrations of gases dissolved within the DIW to equilibrate with the air, which ensured constant concentrations of gases between experiments. Three ml of DIW were transferred to a 50-mm-diameter petri dish inserted in the SDBD discharge chamber and treated for 15 min. The free surface of the water was approximately 3 mm far from the SDBD HV electrode. The PAW was then immediately used to resuspend bacteria to a nal concentration of 10 7 CFU ml − 1 . The LIVE/DEAD ®BacLight™ (Molecular Probes) viability kit was used to assess the viability of bacteria cells treated with PAW. The kit contains a vial of microsphere suspension, and two nucleic acid dyes SYTO 9 and propidium iodide (PI) that allow distinguishing live cells with intact plasma membranes (green) from dead bacteria with compromised membranes (red). Bacteria suspensions were incubated at room temperature for 15 min in the dark in a solution of equal volumes of the two stains. Photomicrographs were taken on a Nikon E800 microscope using a uorescein isothiocyanate (480/30 excitation lter, DM505 dichroic mirror, 535/40 emission lter) and tetramethylrhodamine isocyanate (546/10 excitation lter, DM575 dichroic mirror, 590 emission lter) uorescence lter sets.

Plasma Current Voltage and Emission Spectroscopy
The SDBD was operated in humid ambient air (25°C, 40% RH) at atmospheric pressure. The applied AC voltage (20.6 kVpp (peak-to-peak), see Fig. 1) consisted of repetive burst (repetition rate 500Hz), each consisting of nine AC cycles (f AC = 19.6 kHz) with a repetition rate of 500 Hz (duty cycle of 0.25). In this way we obtained a homogeneous distribution of microdischarges. The actual discharge ON-time is shorter than the duration of the AC cycles. We can estimate an effective duty cicle of the order of 0.1 of the nominal one 53 . This ensures that biological samples are not exposed to excessively damaging temperatures. Typical voltage, current, and charge characteristics are shown in Fig. 1A  setting the rotational temperature to 350 ± 25 K. This temperature represents the gas temperature during the discharge on phase near the dielectric surface, where plasma is con ned in SDBDs. Thus, it does not represent the temperature of the gas in direct contact with the treated substrate. Also, considering that the gas ows at 7 slm, which gives a residence time in the discharge volume of 64 ms, we can rule out heat accumulation within the gap.   The results of EIS and BDS allowed rapid monitoring of PAW properties. The effect of plasma treatment was evaluated by ion density n and estimation of ionic conductivity. In this way, it was possible to predict the effectiveness of the PAW treatment on the XfDD.
In Fig. 6, we show the comparison between NPs (Fig. 6A) as calculated from the impedance of untreated deionized water (DIW) and PAW. The latter refers to two representative plasma treatment times of 5 min and 15 min, which are of interest for this work. The inset shows the magni ed view of the region in the high-frequency region.
It is worth noting that there are signi cant differences between the NPs associated with the DIW (black curves) and the NPs from PAW.
We obtained the best t with the circuit (see Fig. 6B and Table 1) consisting of the series of two components, namely the parallel R 1 //Z CPE related to the response in the low-frequency region and speci cally in the f < 100 Hz region, and a modi ed Randle circuit related to the f > 100 Hz frequency region. The R 1 //Z CPE component is assigned to the interaction of the solution at the interface with the electrodes (electrode polarization, EP effect), while the Randle circuit corresponds to the electrochemical processes within the solution 69 . Following the theory of Bard and Faulkner, the processes involve redox reactions whose effectiveness can be evaluated by the value of the charge transfer resistance R ct , which is also an index of the magnitude of the charge transfer current I ct (I ct ∝R ct −1 ), and then for the ions and free electrons involved in the redox process 70 .
It is noteworthy that the impedance Z R = R ct + Z w as a function of the treatment time τ is related to the concentration of ionic species determined by the optical absorption methods ρ abs (Fig. 6B) and represents a ngerprint of the production and total evolution of NO 3 − and NO 2 − radicals and hydrogen peroxide concentration with treatment time. On the one hand, the decrease of ion diffusivity underlined by the increase of Z W (Z W ∝D − 1 ) is due to the increase of ion concentration; on the other hand, the R ct shows the increase of charge transfer current (I ct ∝R ct −1 ) and then the presence of free redox electrons and underlines the onset of a dominant oxidant environment.
We gain further information on ionic diffusivity D and species concentration n determined via the analysis of dielectric permittivity returned from BDS data. In Fig. 7 we summarize the AC conductivity (left axis) (σ(ω ) vs ω) and the loss factors tan δ frequency dispersion (tan δ vs ω) (right axis) together with the values of the ionic diffusivity, D, and concentration, n, calculated with expr. 5,6,7. Interestingly, the value of the continuous part of the AC conductivity σ DC evidenced by the plateau starting at around f > 100Hz (Fig. 7A) is increasing with the treatment time. Moreover, the σ DC PAW response is not much affected by an ageing time of 24 h (15 min ag) thus con rming the stability in a time of the oxidant environment.
We also compare the tanδ in PAW with that one of the well-known chemical reactions in a solution H 2 O 2 + HNO 3 (1:1) progressively diluted in DIW (data not shown). This is because in both cases the reaction products are the same (i.e., NO 3 − , NO 2 − and H 2 O 2 ). Moreover, the dilution in DIW allowed the use of the known ionic charges' concentration of the chemical reaction to calibrate those observed in PAW vs treatment time (data not shown). The cross-checks of the tanδ of the chemical reaction and PAW spectra allowed us to conclude that BDS results on PAW are consistent with those of the chemical reaction either because of the similar spectral features or the concentration of the ionic charge whose value was found furthermore in agreement with those derived from optical methods.
We can observe (Fig. 7A) that the position tan δ peak of PAW shifts to a higher frequency than that of DIW, which is associated with a decrease in peak intensity compared to treatment time. In addition, the 15min ag PAW item shows a slight increase in tan δ peak intensity with no change in frequency and similar values of σ DC . Consequently, we observed a decrease in the concentration of ionic charges and a corresponding increase in the diffusivity D. This behaviour could be related to the decomposition of hydrogen peroxide and nitrogen peroxyacids 52 . However, the results of EIS and BDS show that the prooxidant environment can still be effective after 24 hours.
Although we can only quantify the total concentration of ionic species at this time, we found a linear correlation between the ion diffusivity D or the total concentration of PAW ionic species n and the total concentration ρ abs (A) measured by chemical methods.

XfDD growth conditions and plasma treatments
Treatment on agar surface-grown bacteria Preliminary investigations showed that the plasma pre-treatment of the agar plates did not alter the growth of XfDD cells (data not showed).
When bacteria (10 7 CFU ml -1 ) were plated out on Buffered Charcoal Yeast Extract (BCYE) agar and exposed to SDBD for 200 s, complete inhibition of cell growth was observed, whereas after 100 s and 10 s of exposure (Fig. 8) there was less pronounced but time-dependent inhibition.
Since it is di cult to quantify the effect of plasma treatment by counting colonies at a concentration of 10 7 CFU ml − 1 (Fig. 9), decimal dilutions from 10 5 to 10 3 CFU ml − 1 of XfDD inoculum were applied to the BCYE agar plates in subsequent experiments (trials). Final SDBD plasma treatment conditions were therefore performed on cells grown 1-5 days after seeding, with a maximum exposure time of 200 s. Table 2 shows the results for different three dilutions (10 3 , 10 4 , 10 5 ) for untreated and 200 s treated plates pre-cultured for one or ve days before LTP exposure. The treatment was effective and complete removed of all bacteria by plasma, at both 1 and 5 days.  The results of these experiments are given in Table 3 and show that in Trial a) a strong effect of cumulative dose was observed, since at the lower concentration (10 3 CFU ml − 1 ) the bacteria were completely killed by the multiple exposures, resulting in a reduction of at least log 5, while at higher CFU the reduction reached at least a log 2 value. In Trial b), cells were counted 20 days after plating. The concentration-dependent effect was less pronounced. Treatment resulted in a reduction of at least log 1 for the lower concentrations (10 3 CFU ml − 1 , 10 4 CFU ml − 1 ), whereas at 10 5 CFU ml − 1 counting was not possible for the untreated sample, making it di cult to estimate the reduction.

Treatment of cells with Plasma Activated Water (PAW)
To test the e cacy of plasma treatment and its potential for use in vivo, bacteria were suspended in PAW previously treated in the SDBD discharge chamber for 15 minutes. To decipher the exact mode of action of the plasma treatment, its effect on XfDD was monitored using viability assays with uorescence live/dead staining.
Fluorescent probes were used to assess cell membrane integrity and XfDD viability after incubation in PAW. The untreated control cells were almost all stained green with SYTO 9, indicating that they were viable. A small number of cells were stained red with propidium iodide (PI), indicating that they were probably dead (Fig. 10A). The number of cells stained red with PI increased dramatically after treatment with PAW, while cells stained with SYTO 9 decreased as expected and only a few uorescent cells were visible after treatment ( Fig. 10 (B)). This indicates that the integrity of the cell membrane of the bacteria was damaged by the treatment, affecting their viability.

Discussion And Conclusions
In the present study, we investigated the effects of LTP application on XfDD inactivation. Our results showed that an exposure time of 200 seconds was su cient to reduce the number of bacterial cells to an undetectable level, proving that LTP is a cost-effective and environmentally friendly alternative to disinfection by chemical and physical means. Furthermore, we demonstrated that the LTP effect is cumulative, opening the possibility of multiple exposure runs.
The plasma chemistry is highly dependent on the composition of the feed gas, the con guration of the system and the operating conditions. Air plasma is a rather complex environment in terms of chemical composition (RONS) and light emission (from UV to mid-range IR, mainly due to molecular nitrogen systems). Using the optical emission spectrum (OES) recorded at about 3 mm from the SDBD surface, we observed weak emissions from the Nitric Oxide gamma band covering the UV-C region (200-280 nm), from OH (mainly in the presence of water) and N2 SPS in the UV-B region (280-315 nm). Thus, direct photooxidation of the protein coat is not considered to be the dominant inactivation process. The strongest emission was observed in the UV-A region (315-400 nm), due to the SPS, FNS band systems of N2. Although we cannot exclude UV-A-mediated bacterial disinfection 71,72 , this cannot be the main process leading to cell death due to the reported exposure time and energy density.
From the SPS (0,0) emission and comparison with the corresponding simulation spectra, we can also rule out heat as a possible mechanism for the disinfection of the samples.
From the emission spectroscopy in the mid-IR, we observed also the presence of atomic oxygen at 777 nm could not be detected. The discharge currently operates in ozone mode, since ozone is the main postdischarge product along with NOx. Thus, the main disinfection mechanism could be due to oxidative pressure by the reactive oxygen and nitrogen species (RONS), which are produced by plasma and are toxic to bacterial pathogens at high concentrations. These RONS oxidize proteins, lipids and nucleic acids and lead to the destruction of the pathogen. They also cause epigenetic regulation that could abrogate bacterial pathogenicity 73 . The e cacy of the treatment xed at 200 s makes disinfection a rather rapid process in the case of planktonic cells. Moreover, pretreatment of the cell culture medium with plasma does not result in su cient changes in substrate chemistry leading to apoptosis of the cells. The presence of bio lm in the cell culture makes it di cult to treat once with the selected xed time. This can be overcome by increasing the plasma dose. Since we wanted to keep the plasma parameters constant, we decided to increase the plasma dose by repeating the treatment with a different frequency. This guarantees that we keep the temperature of the plasma phase close to room temperature and preserve the electrode from possible failure.
In this case, the best results were obtained with a repeated daily treatment for 1 week, which stopped cell growth and resulted in at least a log-2 reduction in bacterial colonies.
We could conclude that the main mechanism of the process is related to the RONS generated by the discharge device. However, for the practical use of LTP, the bacteria must be treated in their natural habitat, the xylem vessels of the plant. To meet these biological requirements, we investigated the use of PAW, to inactivate the bacterium, as this is a valid approach to control XfDD in plants.
PAW has attracted much attention from researchers in the last decade due to its non-thermal and nontoxic mode of action, which is mainly due to the reactive species that can react with the bacterial structural components and later with organelles, proteins and nucleic acid, leading to cell death 74 It has been shown that the effect of the active species produced on PAW on bacteria depends on the bacterial species: In general, gram-negative bacteria were found to be more sensitive to PAW than gram-positive bacteria due to signi cant differences in cell wall structures, physiological state, and ultimately planktonic or bio lm status 75 . The PAW contains RNS and ROS, which can enter bacterial cells through the immediate pores in the active transport cell membrane. The RNS and ROS can oxidize DNA, proteins, and lipids in the cell, breaking DNA, degrading proteins, and inducing lipid peroxidation thereby causing the contents to ow out of the bacteria and die 59 , with ozone and peroxynitrite being the dominant species in the sterilization process 53,60−62 . Many researchers also speculate that acidity and active compounds are related in PAW. In an acidic environment, the RNS and ROS will react with the lipids and carbohydrates of the DNA proteins in the cells, lowering the pH of the cells and causing physiological dysfunction and cell death63. In addition, the effects of PAW on bio lms have also been explored [75][76][77][78] .
The importance of biochemically reactive species formed in PAW for the destruction and degradation of the bio lm matrix and the release of resident microbial cells have been highlighted 79 .
In our experiment, we used freshly prepared PAW obtained by activation of sterile water by SBDB in air.
Assuming that the activity of PAW in terms of sterilisation properties is due to the presence of peroxynitrous acid (which requires nitrite, peroxide and acidic conditions) 52 , freshly prepared PAW by cold plasma processes should be the most e cient means. Chemical analysis of freshly prepared PAW at a treatment time of 15 minutes showed the presence of nitrite, nitrate, and hydrogen peroxide. Nevertheless, the PAW proved to be su ciently stable for 1 day when stored in an opaque container at a temperature of 4°C. The EIS and BDS were used as rapid and non-destructive techniques to monitor the redox activity of PAW as a function of plasma treatment time. This is evidenced by the behaviour of the charge transfer resistance and DC conductivity, both of which are markers of a pro-oxidant environment. In our case, the latter activity is particularly pronounced at the 15-min treatment time PAW as can be deduced from the values of both electrical parameters. Interestingly, we found a correlation between the total ionic charge concentration and diffusivity determined by EIS and DBS and the total ionic concentration determined by optical methods. The higher nitrate, nitrite and peroxide content of the freshly produced PAW was indicative of the best treatment e cacy in this case. Nevertheless, the aged PAW redox environment is an indication that the treatment could be effective even after 24 hours.
Our results showed that PAW has excellent antimicrobial potential to inactivate X. fastidiosa cells. Just 15 minutes of treatment is su cient to destroy XfDD cells in in vitro experiments. This is an important step towards the development of plasma-assisted strategies to inhibit the growth or kill XfDD in the xylem vessels of plants and to apply an environmentally safe strategy to control this pathogen, which will be the progress of the present work.      the plateau at the angular frequency max (C) Correlation between ionic species obtained via optical absorption methods abs and BDS data, n.   Fluorescence images of XfDD suspension in water for the untreated (A), and PAW treated (B) samples.