Antimicrobial Peptides With Antibiofilm Activity Against Xylella fastidiosa

Xylella fastidiosa is a plant pathogen that was recently introduced in Europe and is causing havoc to its agriculture. This Gram-negative bacterium invades the host xylem, multiplies, and forms biofilm occluding the vessels and killing its host. In spite of the great research effort, there is no method that effectively prevents or cures hosts from infections. The main control strategies up to now are eradication, vector control, and pathogen-free plant material. Antimicrobial peptides have arisen as promising candidates to combat this bacterium due to their broad spectrum of activity and low environmental impact. In this work, peptides previously reported in the literature and newly designed analogs were studied for its bactericidal and antibiofilm activity against X. fastidiosa. Also, their hemolytic activity and effect on tobacco leaves when infiltrated were determined. To assess the activity of peptides, the strain IVIA 5387.2 with moderate growth, able to produce biofilm and susceptible to antimicrobial peptides, was selected among six representative strains found in the Mediterranean area (DD1, CFBP 8173, Temecula, IVIA 5387.2, IVIA 5770, and IVIA 5901.2). Two interesting groups of peptides were identified with bactericidal and/or antibiofilm activity and low-moderate toxicity. The peptides 1036 and RIJK2 with dual (bactericidal–antibiofilm) activity against the pathogen and moderate toxicity stand out as the best candidates to control X. fastidiosa diseases. Nevertheless, peptides with only antibiofilm activity and low toxicity are also promising agents as they could prevent the occlusion of xylem vessels caused by the pathogen. The present work contributes to provide novel compounds with antimicrobial and antibiofilm activity that could lead to the development of new treatments against diseases caused by X. fastidiosa.


INTRODUCTION
Xylella fastidiosa is a Gram-negative xylem-inhabiting bacterium that causes important plant diseases that pose great threats to the agriculture worldwide (Purcell, 2013). This pathogen was first detected in California in grapevines causing Pierce's disease (Alston et al., 2013). It is also responsible for other plant diseases such as citrus variegated chlorosis (Rapicavoli et al., 2018) and almond leaf scorch disease. In 2013, it was introduced in Italy and is spreading through the Mediterranean region causing a new disease named olive quick decline syndrome (EFSA, 2013). The increasing dissemination of X. fastidiosa can be related to many factors, such as climate conditions optimal for its growth, its easy spread through insect vectors from the Cicadellidae (sharpshooter leafhoppers) or the Aphrophodridae (meadow spittlebug) families, and the huge number of hosts that it can infect (Almeida and Nunney, 2015;Strona et al., 2017). Therefore, this pathogen could cause havoc in the agricultural economy of countries that are important global producers of olives, citrus, almonds, and grapes, such as Italy, Spain, France, and Greece (Food and Agriculture Organization of the United Nations, 2019).
Since X. fastidiosa inhabits xylem vessels in host plants, biofilm formation is the main pathogenic mechanism for the symptomatology of plants infected by this pathogen (Cardinale et al., 2018). Once X. fastidiosa is inoculated into the host xylem vessels by an insect vector, the cells first remain in a planktonic stage and then are reversibly attached to the vessels' surface. Next, cells are irreversibly embedded in a self-produced matrix of extracellular polymeric substances (EPS) leading to the formation of the biofilm (Cattò et al., 2019). Eventually, the architecture of this biofilm matures and reaches its maximum complexity occluding the xylem vessels, blocking the sap flow and depriving the plants of water and nutrition (Martelli et al., 2016). Finally, cells detach from the biofilm and become planktonic again, being able to disperse to other areas of the plant (Mendes et al., 2016). In this planktonic state, cells can be acquired by vectors when they feed upon the xylem of infected plants spreading the pathogen to healthy plants.
At present, most of the measures adopted to manage the diseases caused by X. fastidiosa are aimed to limit the spread of the bacterium. Some of these strategies are related to agricultural practices such as the application of insecticides to control the vector population and the eradication of infected plants (EFSA, 2016). Europe is migrating to a more sustainable agriculture model so many chemical compounds used in the past to control bacterial plant pathogens have been prohibited or restricted to be used on field (Navarrete and De La Fuente, 2014;ECDC, EFSA and EMA, 2015;EFSA, 2016). Nevertheless, different approaches have been studied consisting of new chemicals and biological control strategies. Some chemical compounds such as N-acetyl-L-cysteine (NAC) in citrus plants (Muranaka et al., 2013), copper (II) sulfate in tobacco plants (Ge et al., 2020), and menadione, benzethonium chloride, and abscisic acid in grapevines (Meyer and Kirkpatrick, 2011;Zhang et al., 2019) seem to be effective in greenhouse conditions. Moreover, the antibiotic oxytetracycline along with three other compounds, like NAC, a bioactive detergent composed of plant oil extracts, and a Zn, Cu, and citric acid fertilizer, showed potential to be used to control X. fastidiosa diseases in almond (Amanifar et al., 2016) and olive orchards (Dongiovanni et al., 2017;Scortichini et al., 2018;Bruno et al., 2021), respectively. Other strategies that have been studied involve the use of the endophyte Paraburkholderia phytofirmans (Baccari et al., 2019), avirulent X. fastidiosa strains (Hao et al., 2017), and lytic phages (Das et al., 2015) as biological control agents. Although the results obtained in most of these trials were positive, no strategy was able to completely cure plants infected by X. fastidiosa. Therefore, there is still a need to find efficient compounds and eco-friendly alternatives that comply with the European environmental regulations.
Antimicrobial peptides are a class of peptides that could be considered as promising candidates to control X. fastidiosa. In general, they exhibit high antibacterial activity and low toxicity (Guell et al., 2011;Li et al., 2020;Liang et al., 2020). In addition, they are not persistent compounds and resistance to them in pathogens is difficult to emerge since their mechanism of action mainly involves cell membrane disruption (Yeaman, 2003;Brogden, 2005;Peschel and Sahl, 2006;Von Borowski et al., 2018). Up to now, few antimicrobial peptides with activity against X. fastidiosa have been reported. In particular, indolicidin and magainin 2 have shown activity against several strains with minimum inhibitory concentration (MIC) between 8 and 64 µM (Li and Gray, 2003;Kuzina et al., 2006;Fogaça et al., 2010). Moreover, we recently identified the bactericidal peptides BP171 and BP178, which are active against several X. fastidiosa strains with a reduction in viability approximately 3.6 log at 12.5 µM (Baró et al., 2020a,b).
It is worth mentioning that, despite the fact that biofilm is the main virulence factor for X. fastidiosa, there have not been reported peptides able to inhibit its formation. At the moment, only a few non-peptidic compounds have been reported to present some antibiofilm activity against this pathogen such as the previously mentioned, NAC and the Zn, Cu, and citric acid fertilizer, DOX-derived oxylipins, and phenolic compounds such as gallic acid and epicatechin (Muranaka et al., 2013;Lee et al., 2020;Scala et al., 2020;Tatulli et al., 2021). Nevertheless, peptides with antibiofilm activity against other Gram-negative bacteria (Escherichia coli, Pseudomonas aeruginosa, Acinetobacter baumannii, Klebsiella pneumoniae, and some species of Salmonella) or sequences with both antibacterial and antibiofilm activity have been widely described. These peptides could be considered good candidates to be tested against X. fastidiosa. Among these potential candidates, peptides from the family of RR showed antibacterial and antibiofilm activity against multidrug resistant clinical strains (Mohamed et al., 2017). Other peptides that displayed both antibacterial and antibiofilm activity are the LL-37 derivative KR-12-a5 and the peptide SB056 (Batoni et al., 2016;Kim et al., 2017). De La Fuente-Núñez et al. described antimicrobial peptides that target biofilm formation, including LJK2 and its retro-inverso analog RIJK2, and the innate defense regulator IDR-1018 (De La Fuente-Núñez et al., 2015). These authors also identified the small cationic antimicrobial peptide HH15 and its analogs 1026, 1029, 1036, and 1037, which displayed antibacterial and/or antibiofilm activities (De La Fuente-Núñez et al., 2012). All these peptides share the consensus sequence FRIRVRV-NH 2 (FV7), which was later proven to be active and used to design the conjugate R-FV7-I16 (Xu et al., 2014). Scorpion venom peptides AamAP1 and HP1404 have also been described to display interesting biological properties and their sequence has served as basis for the design of new analogs, including AamAP-S1, HP1404-T1D, and HP1404-T1E (Almaaytah et al., 2012;Kim et al., 2018).
Based on these considerations, the aim of the present work was to identify peptides able to control X. fastidiosa. First, the differential susceptibility of X. fastidiosa strains to antimicrobial peptides was assessed in order to select a representative strain to evaluate the activity of the peptides. Then, we synthesized the peptides mentioned above together with several new analogs, and tested them for their bactericidal and antibiofilm activity against X. fastidiosa. In addition, their effect on leaf infiltration in a tobacco plant model and their hemolytic activity were studied.

Synthesis of Peptides
Peptides (Table 1) were synthesized manually on solid phase using standard 9-fluorenylmethoxycarbonyl (Fmoc)/tertbutyl (tBu) strategy. A Fmoc-Rink-ChemMatrix resin (0.69 mmol/g), a PAC-MBHA resin (0.24 mmol/g), or a Fmoc-Rink-MBHA resin (0.56 mmol/g) was used as solid support. The Fmoc-Rink-ChemMatrix resin was selected for the synthesis of peptides containing more than 14 residues. The PAC-MBHA resin was employed to prepare C-terminal carboxylic acid peptides whereas the Fmoc-Rink-ChemMatrix and the Fmoc-Rink-MBHA resins served for C-terminal peptide amides. Peptide elongation was carried out through sequential steps of Fmoc removal and coupling of the corresponding amino acid as previously described (Caravaca-Fuentes et al., 2021;Oliveras et al., 2021). Once the peptide sequence was completed, each resulting peptidyl resin was treated with trifluoroacetic acid (TFA)/H 2 O/triisopropylsilane (TIS) (95:2.5:2.5). Peptidyl resins that contained tryptophan and/or arginines were treated with TFA/H 2 O/TIS/thioanisole/1,2-ethandithiol/phenol (81.5:5:1:5:2.5:5). Following TFA evaporation and diethyl ether extraction, the crude peptides were purified by reverse-phase column chromatography, lyophilized, analyzed by HPLC, and characterized by mass spectrometry.  Table 2). All strains were stored in Pierce disease broth (PD2, Davis, 1980) supplemented with glycerol (30%) and maintained at -80 • C. When needed, strains were cultured in buffered charcoal yeast extract (BCYE) agar plates (Wells et al., 1981) at 28 • C for 7 days. Afterward, colonies were scrapped and cultured in new BCYE media at 28 • C for 7 additional days before being used in any of the experiments. When liquid cultures were required, PD3 broth (Davis et al., 1981) was used. Cell suspensions were prepared in sterile succinate-citrate-phosphate buffer (SCP) for bactericidal experiments or in sterile phosphate-buffered saline buffer (PBS) for biofilm experiments. The suspensions were adjusted to an optical density at 600 nm (OD 600 ) of 0.32, which corresponds approximately to 10 8 CFU/ml, which was confirmed by plate counting in PD2 modified with Gelrite TM (9 g/l). Growth curves of selected X. fastidiosa strains were performed by culturing a cell suspension prepared in 180 µl of PD3 medium (adjusted at an OD 600 of 0.1) and mixed with 20 µl of water in 96-well plates (Nuclon TM Delta Surface, Thermo Fisher Scientific, Spain). Three replicates of 10 wells were prepared for each strain. The microplates were incubated at 28 • C under shaking (120 rpm) for 7 days and measures of OD 600 were performed each day using the EPOCH2 TC microplate reader (BioTek, Winooski, United States). Background values of OD were subtracted from data and area under the growth curve (AUGC), specific growth rate, and doubling time were calculated for each replicate and strain.
Biofilm formation was quantified at the end of the growth curve experiment described above using the crystal violet dye according to the methods previously described (Zaini et al., 2009). The total growth, planktonic growth (cells in suspension), and biofilm formed (cells adhered to the well surface) were estimated by measuring OD. Planktonic cells were recovered from the media and transferred into new microplates and OD 600 was measured. To quantify the biofilm formed, the original 96well plate was rinsed gently with sterile distilled water three times, stained with 250 µl of crystal violet (0.1%) for 20 min, and rinsed with sterile distilled water three times to discard excess dye. Finally, crystal violet adhered to the biofilm was solubilized with 250 µl of a mixture of ethanol/acetone (4:6) for 10 min and a measure of OD 595 was made. Two independent experiments of the biofilm formation were carried out with three replicates of 10 wells for each strain.
Time course of biofilm formation was assessed in order to select the best time for biofilm formation of Xff IVIA 5387.2. Different times of incubation (from 1 to 7 days) were tested. In each experiment, growth, planktonic cells, and biofilm formation were measured after the selected incubation period as described above.

Bactericidal Activity
Bactericidal activity of the peptides was assessed by a test contact coupled with viable-quantitative PCR (v-qPCR) as previously described (Baró et al., 2020a). Sensitivity and amplification efficiency of the v-qPCR were evaluated for all studied strains. Briefly, standard curves were prepared using viable, dead (by heating them at 95 • C for 20 min), or a mixture of viable and dead cells. Dilutions of a homogeneous cell suspension in SCP buffer (from 10 8 to 5 × 10 2 CFU/ml) of viable or dead cells to a total volume of 200 µl in DNA low binding tubes were prepared. Mixtures of viable cells with a constant number of dead cells (1 × 10 6 CFU/ml) were also included to assess the influence of dead cells. Two sets of dilutions for viable, dead, or mixture were prepared and one of them was treated with PMAxx (VWR, Barcelona, Spain). Briefly, PMAxx was added at a final concentration of 7.5 µM, and samples were incubated for 8 min in the dark at room temperature following a photoactivation of 15 min (PMA-Lite TM LED Photolysis Device, Biotium, CA, United States). DNA extractions of all samples were performed using the GeneJET Genomic DNA Purification Kit (Thermo Fisher Scientific, United States) following the specific protocol for Gram-negative bacterial suspensions and were analyzed in duplicate by a TaqMan-based qPCR assay based on the 16S rRNA sequence (Baró et al., 2020a). Then, a calibration curve for each strain with and without PMAxx was calculated by using cell concentration and C T values, determined by qPCR. Three independent experiments were performed for each curve. X. fastidiosa strains' susceptibility to the peptide BP171 was tested by a contact exposure test combined with v-qPCR against the six X. fastidiosa strains (Table 2) as previously described (Baró et al., 2020a). Briefly, the peptide was solubilized in sterile Milli-Q water to a stock concentration of 1 mM and filter sterilized through a 0.22 µM pore size filter. BP171 was tested at a final concentration of 3.1 and 12.5 µM. Twenty microliters of the corresponding peptide dilution were mixed with 180 µl of a X. fastidiosa suspension, as described above. Three biological replicates for each concentration were performed and a nontreated control with sterile water instead of the peptide was included. Contact tests were incubated at room temperature for 3 h. Afterward, each tube was treated with PMAxx and was handled as previously described. The reduction in viability, expressed as log 10 CFU/ml, was obtained by interpolating the C T values from each sample against the respective standard curve for each strain and subtracting it from the non-treated control (Log 10 (N 0 /N)).
The bactericidal activity of the selected peptides ( Table 1) at 50 µM against Xff IVIA 5387.2 was determined as described above. Cecropin B (C1796, Merck, Spain) was also tested as reference control (Li and Gray, 2003). Highly active peptides (reduction in viability > 3 logs) were further tested at 12.5 and 3.1 µM to better characterize their bactericidal activity.

Antibiofilm Activity
The effect of N-acetyl-L-cysteine (NAC; A9165, Merck, Spain) on biofilm formation of the studied strains was determined since it was previously described to reduce biofilm formation of X. fastidiosa (Muranaka et al., 2013). NAC was tested at a final concentration of 50 µM. Twenty microliters of NAC were mixed with 180 µl of a X. fastidiosa suspension in PD3 in 96-well plates, as previously described in this study. Three replicates of 10 wells were made for each strain. Microplates were incubated at 28 • C for 5 days under continuous shaking (120 rpm). Finally, growth, planktonic cells and biofilm formation were measured as previously described. The ratio of biofilm formation was calculated according to the formula Oi/Oc, where Oi is the OD 595 of the treatment and Oc is the OD 595 of the non-treated control. The ratio of planktonic cells was calculated as described above but OD was measured at 600 nm.
To assess the antibiofilm activity of all the synthetized peptides, they were prepared as described in the bactericidal activity experiments. They were tested for antibiofilm activity at a final concentration of 50 µM against Xff IVIA 5387.2 as previously described in this study. Magainin 2 was tested at 12.5 µM and RIJK2 and 1036 were tested at 3.1 µM to prevent the influence of their antimicrobial activity in the biofilm formation.
To analyze the effect of peptides 1026, RJK2, and R-FV7-I16 in biofilm detachment of Xff IVIA 5387.2, a quantification of cells by qPCR including the biofilm attached, biofilm detached, and planktonic cells was carried out. One hundred microliters of the peptides 1026, RJK2, or R-FV7-I16 were mixed with 900 µl of a X. fastidiosa suspension to a final peptide concentration of 50 µM in each well of a 24-well microplate. Non-treated wells were included as controls by substituting the volume of peptide with sterile water. A total of three replicates were made for treatment in each experiment. Two independent experiments were performed. Microplates were incubated at 28 • C for 5 days under continuous shaking (120 rpm). Planktonic cells were recovered into tubes and centrifuged at 13,000 rpm for 10 min. Biofilm detached cells were recovered from the rinsing water by transferring the content of each well into a tube and centrifuging the mixture at 13,000 rpm for 10 min. This operation was repeated a total of six times and all the washes were collected in the same tube. Biofilm attached cells were recovered from each well by adding 1 ml of PBS, scrapping the attached cells with an inoculation loop, transferring them into a tube, and centrifuging them at 13,000 rpm for 10 min. All the pellets were suspended with PBS to a total volume of 1 ml. DNA extraction was performed for each sample and DNA samples were analyzed in duplicate by a TaqMan-based qPCR as previously described in the bactericidal activity experiments of this study.
Dose-effect relationship of BP525, 1037, and R-FV7-I16 on biofilm inhibition was determined. They were tested at 0, 6.3, 12.5, 25, and 50 µM against Xff IVIA 5387.2 as described in this study. Three replicates of 10 wells were made for each peptide and concentration. For dose-response modeling in inhibition of biofilm formation, percentage of biofilm inhibition (B i ) was calculated according to the formula: Oi is the OD 595 of the treatment and Oc is the OD 595 of the non-treated control.

Effect of Peptide Infiltration on Tobacco Leaves
Peptides were evaluated for their effect upon infiltration on tobacco leaves as previously described (Nadal et al., 2012). Briefly, tobacco plants (Nicotiana tabacum) were grown from seed in a heated glasshouse and used between 20 and 30 days old. Using a syringe, 100 µl of peptide solutions of 50, 100, and 150 µM were infiltrated into the mesophyll of fully expanded tobacco leaves (previously wounded with a needle). Six independent inoculations were carried out in a single leaf, and three independent inoculations were performed per peptide and concentration randomly distributed in different leaves and plants. Control infiltrations with water (negative control) or melittin (M2272, Merck, Madrid, Spain) (positive control) at the same molar concentrations were performed. Plants were kept at standard greenhouse conditions for 48 h. Peptide's leaf infiltration effect was measured as the lesion diameter.

Hemolytic Activity
The hemolytic activity of peptides was used as an indication of its toxicity, according to the current literature in this field Inui Kishi et al., 2018). It was assessed by determining hemoglobin release from erythrocyte suspensions of horse blood (5% vol/vol) (SR0050C, Thermo Fisher Scientific, Spain) as previously described (Badosa et al., 2007). Briefly, peptides were solubilized in TRIS buffer and mixed with cleaned 10-fold diluted horse erythrocytes. The final peptide concentrations tested were 150, 250, and 375 µM. The percentage of hemolysis (H) was calculated using the equation: where Op is the optical density at 540 nm for a given peptide concentration, Ob for the buffer, and Om for the melittin positive control.

Data Analysis
Specific growth rates were estimated based on the slope of the growth curve (ln OD 600 vs. time) at the exponential phase (Supplementary Figure 1). They were determined between 1 and 3 days for IVIA 5387.2 and Temecula strains, between 1 and 4 days for IVIA 5770 and CFBP 8173 strains, and between 3 and 6 days for DD1 and IVIA 5901.2 strains. The doubling time for each strain was calculated using the formula ln2/specific growth rate. To test the significance of the effect of strain on the parameters presented in Table 3, a one-way analysis of variance (ANOVA) was used. In all cases, means were separated according to the Duncan's test at a p-value of < 0.05 (IBM SPSS Statistics for Windows, Version 25.0 released on 2017 by IBM Corp, Armonk, NY, United States).
Also, to test the significance of the effect of peptides, peptide concentration, and time in the experiments, one-way ANOVA was performed. In all cases, means were separated according to the Duncan's test (p < 0.05).
Data on peptide dose-biofilm inhibition were adjusted to a Michaelis-Menten model to determine the maximum biofilm inhibition (B i max) and the median effective dose (ED 50 ): where, a is the B i max and b is the ED 50 (Waghu et al., 2018).
Principal components analysis (PCA) was used to evaluate singularities among the tested peptides to select the ones with the best biological profile (IBM SPSS Statistics for Windows, Version 25.0 released on 2017 by IBM Corp, Armonk, NY, United States). PCA was performed using 31 peptides on five variables: (i) bactericidal activity as the reduction in viability of Xff IVIA 5387.2, (ii) antibiofilm activity as the ratio of biofilm formation of Xff IVIA 5387.2, (iii) planktonic cells as the ratio of planktonic cells of Xff present after the peptide treatment, (iv) hemolytic activity as the percentage of hemoglobin release from erythrocyte suspensions of horse blood, and (v) leaf infiltration effect as the lesion diameter on tobacco leaves.
These 31 peptides were manually synthesized following a standard Fmoc/ t Bu strategy. They were obtained in excellent  4 The reduction in viability was calculated as Log N 0 /N where N 0 is 10 7 CFU/ml of a non-treated control and N is CFU/ml of the treatment. Values are the means of three replicates plus the confidence interval (α = 0.05). Means sharing the same letters are not significantly different (p < 0.05) according to the Duncan's test. 5 All values are represented as a ratio between the OD 595 obtained after the treatment (T) and the OD 595 of a non-treated control (NTC). Values are the means of three replicates of 10 wells plus the confidence interval (α = 0.05). Means sharing the same letters are not significantly different (p < 0.05) according to the Duncan's test. HPLC purities (93->99%), except for magainin 2 (64%), and their identity was confirmed by mass spectrometry (Supplementary Table 1).

Growth and Biofilm Formation of Xylella fastidiosa Strains
Six X. fastidiosa strains belonging to three subspecies were characterized in relation to their capacity for growth and biofilm formation ( Table 3 and Supplementary Figure 1) Growth and biofilm formation kinetics of Xff IVIA 5387.2 was assessed (Figure 1). Total growth (including biofilm and planktonic cells) was characterized by a first stage of linear increase until the fourth day, followed by a stationary phase. Biofilm formation increased until a maximum at the fourth day and then it started to decrease on the seventh. Planktonic cells grew monotonically for the whole experiment. Interestingly, the highest values of planktonic cells were achieved with the decrease in biofilm during the stationary phase of growth.

Bactericidal Susceptibility to Peptides and Effect of N-acetyl-L-cysteine on Biofilm Formation in Xylella fastidiosa Strains
Suitability of the v-qPCR method to quantify viable cells of different X. fastidiosa strains was analyzed. Standard curves obtained for the strains showed efficiencies ranging from 81 to 98.4% and the method had enough sensitivity to detect a minimum of 10 3 CFU/ml of viable cells when mixed with dead cells (Supplementary Table 2 and Supplementary Figure 2).
Then, the susceptibility of six X. fastidiosa strains to the peptide BP171 was assessed at 3.1 and 12.5 µM ( Table 3) using the v-qPCR method. The bactericidal activity of the peptide was clearly dependent on the strains. Globally, Xfm CFBP 8173 displayed the highest reduction in viability followed closely by Xfm IVIA 5901.2 and Xfp DD1 was the most resistant strain to the peptide at both concentrations. IVIA 5387.2, IVIA 5770, and Temecula showed an intermediate resistance.
The effect of NAC on biofilm formation of X. fastidiosa strains was also assessed ( Table 3). The effect was measured as a ratio between the OD 595 values of treated and non-treated wells. NAC treatment affected the biofilm formation of Xff IVIA 5387.2 and Xfp DD1 with a ratio of 0.72 and 0.42, respectively. In contrast, it did not affect significantly the strains Xff Temecula, Xfm CFBP 8173, Xff IVIA 5770, and Xfm IVIA 5901.2 that showed ratios ranging from 0.94 and 1. Considering all of the above, Xff IVIA 5387.2 was selected in subsequent experiments as it showed intermediate susceptibility to the peptides.

Bactericidal Activity
The bactericidal activity of the 31 peptides was tested at 50 µM against Xff IVIA 5387.2 (Figure 2). Peptides were classified into five statistically different groups (Supplementary Table 3). RIJK2, 1036, magainin 2, and the reference peptide cecropin B were highly active, leading to more than 3 log reduction of cell viability. RR4-NH 2 , AamAP-S1, and indolicidin exhibited high activity with 2 to 3 log reduction of cell viability. Seven peptides showed moderate activity with 1 to 2 log reduction of cell viability. Twelve peptides had low activity with a 0.3 and 1 log reduction and seven peptides showed very low activity with less than 0.3 log.
Peptides that showed very high bactericidal activity against X. fastidiosa were further tested at lower concentrations, 12.5 and 3.1 µM (Supplementary Table 3). At 12.5 µM, 1036 was the most active peptide with a higher log reduction than cecropin B (3.48 vs. 3.19). At this concentration, RIJK2 and magainin 2 displayed similar activity with 2.34 and 2 log reduction, respectively. At FIGURE 2 | Bactericidal and antibiofilm activity against X. fastidiosa subsp. fastidiosa IVIA 5387.2, tobacco leaf infiltration effect, and hemolytic activity of peptides. Values are the means of three replicates and error bars represent the confidence interval (α = 0.05). Each color represents a different peptide family. The asterisk (*) indicates the peptides that have the best values for each activity according to the Duncan's test (p < 0.05). The reference compounds used were cecropin B for bactericidal activity, NAC for antibiofilm activity, and melittin for tobacco leaf infiltration effect and hemolytic activity. For antibiofilm activity, peptides with very high bactericidal activity were diluted (magainin 2 at 12.5 µM and RIJK2 and 1036 at 3.1 µM) and LJK2, KR-12-a5, SB056, AamAP1, AamAP-S1, and AamAP-R were not included (ni) since they affected X. fastidiosa's growth.
3.1 µM, except for magainin 2 that was poorly active, the other peptides exhibited log reductions ranging from 1.89 to 2.13. In contrast with the other peptides, RIJK2 was similarly active at both concentrations.

Antibiofilm Activity
The antibiofilm activity of 31 peptides was determined against Xff IVIA 5387.2 at 50 µM (Figures 2, 3 and Supplementary  Table 3). RIJK2 and 1036 were tested at 3.1 µM and magainin 2 was tested at 12.5 mM to minimize the influence of their antimicrobial activity in the biofilm formation. Peptides that affected X. fastidiosa's growth were not included in Figures 2, 3, specifically LJK2, KR-12-a5, SB056, AamAP1, AamAP-S1, and AamAP-R. Peptides were classified into three major groups according to their antibiofilm activity. Twenty out of 25 peptides exhibited high antibiofilm activity (ratio of biofilm formation from 0.01 to 0.2), peptide 1026 being the most active. It is worth mentioning that peptides 1036 and RIJK2 showed a high antibiofilm activity despite being tested at 3.1 µM (ratio of biofilm formation of 0.07 and 0.12, respectively). Three peptides had moderate activity (ratio from 0.29 to 0.47) and two displayed low activity (ratio from 0.72 to 0.77) (Supplementary Table 3).
Nevertheless, peptides with high antibiofilm activity exhibited a high variability in the ratio of planktonic cells that ranged from 0.46 to 2.78 (Figure 3). Peptides 1026, RJK2, and BP527, which showed a high antibiofilm activity but exhibited a different ratio of planktonic cells, were further tested at 50 µM for their effect on biofilm formation of Xff IVIA 5387.2 (Figure 4). Planktonic, biofilm attached, and biofilm detached cells were quantified with qPCR. Biofilm detached cells correspond to a transition between biofilm attached cells and planktonic cells. A total amount of 1.6 × 10 8 to 1.7 × 10 8 CFU/ml of Xff IVIA 5387.2 were quantified in the wells in all treatments after 5 days of incubation. In non-treated cells, 35.9% of biofilm attached cells was observed together with only 2.8% of biofilm detached cells and 61.3% of planktonic cells. The exposure of X. fastidiosa to peptides reduced the biofilm formation and influenced biofilm attachment. In the treatment with peptide 1026, only 0.7% of attached biofilm cells was observed, 19.8% were biofilm detached cells, and 79.5% planktonic cells. RJK2 had a similar activity to 1026. In contrast, BP527 had less antibiofilm activity with 12.5% of attached biofilm cells, 45.7% of detached biofilm cells, and only 41.7% of planktonic cells observed. Therefore, the peptides FIGURE 3 | Effect of peptides on biofilm formation and planktonic cells on X. fastidiosa subsp. fastidiosa IVIA 5387.2. Peptide concentrations were 50 µM except for magainin 2 (12.5 µM), and 1036 and RIJK2 (3.1 µM) which are marked with an asterisk (*). NAC was used as reference. Each color within the symbols represents a different peptide family. Values are the means of three replicates of 10 wells. Peptides were classified by antibiofilm activity using the Duncan's test (p < 0.05) and were labeled as high, moderate, and low antibiofilm activity.
Frontiers in Microbiology | www.frontiersin.org differentially affected the balance between the attachment and detachment of biofilm cells.

Leaf Infiltration Effect on Tobacco Plants and Hemolytic Activity
The effect of the peptides on eukaryotic cells was assessed on tobacco leaves and erythrocytes. The peptide's leaf infiltration effect was determined by infiltrating them into the mesophyll of tobacco plant leaves at 50, 100, and 150 µM (Supplementary Table 4). Melittin was used as a reference. Lesion diameter at 100 µM is shown in Figure 2. Melittin caused the highest lesion (14.7 mm), and except for AamAP1 and BP528, peptides caused a lesion ranging from 0 to 11 mm, which was significantly lower than melittin.
Hemolytic activity of the peptides was determined on erythrocytes and compared to the reference peptide melittin (Supplementary Table 4). Percent hemolysis at 250 µM is shown in Figure 2. Fifteen out of the 31 peptides analyzed showed a hemolysis ≤14% and 5 exhibited a hemolysis between 20 and 44%.

Grouping Peptides According to Their Biological Profile
Five variables were selected for the biological profile analysis of the peptides (bactericidal activity, antibiofilm activity, planktonic cell presence, hemolytic activity, and leaf infiltration effect) to group peptides with a PCA. The first three principal components (PCs) accounted for 48.5, 20.5, and 17.7%, respectively, of the total variation in the dataset. Therefore, the three-dimensional scatter plot of the peptides is a good approximation as it represents 86.7% of the total variation of the data (Figure 6). The PC1 axis represents the variables leaf infiltration effect and hemolytic activity. The PC2 axis reflects the antibiofilm activity. The PC3 axis represents the bactericidal activity. Less toxic peptides have low values in PC1, peptides with higher antibiofilm activity have low values in PC2, and highly bactericidal peptides have high values in PC3.
In the PCA, four major groups and an outlier were identified. The first group was composed of 1036, RIJK2, and magainin 2, which have high bactericidal activity, high antibiofilm activity, and moderate toxicity. The second group was formed by RR4-NH 2 , indolicidin, and AamAP-S1, which had moderate bactericidal activity, high antibiofilm activity, and moderate to high toxicity. The third group was represented by KR-12-a5, AamAP1, AamAP-R, BP527, and HP1404 and had low bactericidal activity, high antibiofilm activity, and high toxicity. The fourth group was composed of all the other peptides except for magainin 2(1-10) and had low/moderate bactericidal activity, high/moderate antibiofilm activity, and low toxicity. Magainin 2(1-10) behaves differently from all the other peptides and it was considered as an outlier.

DISCUSSION
X. fastidiosa is a highly relevant plant-pathogenic bacterium in the European Union due to the high field productivity losses that it causes, which may dampen the local economy (Ferguson et al., 2017;EPPO, 2019;Schneider et al., 2020). Its main mechanism of pathogenicity is biofilm formation that may lead to the host death. Due to the impact of this pathogen, many strategies have been researched to control the diseases caused by X. fastidiosa. In general, promising results were obtained in reducing disease severity but no strategy was able to completely cure infected plants (Amanifar et al., 2016;Dongiovanni et al., 2017;Scortichini et al., 2018;Bruno et al., 2021). In this context, antimicrobial peptides, such as cecropin B, magainin 2, indolicidin, and BP178, have been previously reported to display antibacterial activity against X. fastidiosa (Li and Gray, 2003;Kuzina et al., 2006;Baró et al., 2020a,b). Although the number of reported antimicrobial peptides active against this bacterium are scarce, these examples pave the way to search for new candidates. Moreover, peptide sequences with antibiofilm activity against this pathogen have not been reported. Nevertheless, peptides able to affect biofilm formation have been described against other Gram-negative bacteria, and these sequences could be considered as potential candidates to be tested against X. fastidiosa. Thus, in this paper, the above sequences were taken as the basis for the design and identification of new peptides with bactericidal or antibiofilm activity against X. fastidiosa.
To assess the activity of peptides against X. fastidiosa, a strain displaying moderate values regarding growth, biofilm formation, and susceptibility to antimicrobial peptides was selected among six strains (Xfp DD1, Xfm CFBP 8173, Xff Temecula, Xff IVIA 5387.2, Xff IVIA 5770, and Xfm IVIA 5901.2). These strains belong to three of the major subspecies found in the Mediterranean area (pauca, fastidiosa and multiplex). These subspecies are more or less specific to a particular host range and climate conditions, so it would be expected that they displayed different behaviors in growth, biofilm formation, and susceptibility to antimicrobial compounds as it has been previously reported (Baldi and La Porta, 2017;Denancé et al., 2017Denancé et al., , 2019. In the present study, the strains differed greatly in all the evaluated parameters accordingly to other studies, which also observed noticeable differences between other X. fastidiosa strains regarding growth and biofilm formation (Feil and Purcell, 2001;Guilhabert and Kirkpatrick, 2005) and susceptibility to antimicrobial peptides (Baró et al., 2020a,b). Specifically, Xff Temecula, Xff IVIA 5387.2, and Xff IVIA 5770 displayed moderate values in growth, biofilm formation, and susceptibility to antimicrobial peptides. Interestingly, concerning these three strains, NAC, which was previously reported as an antibiofilm compound (Muranaka et al., 2013), only affected the biofilm formation of Xff IVIA 5387.2. In the case of Xff Temecula, the values of the growth parameters were similar to those previously reported in the literature (Guilhabert and Kirkpatrick, 2005;Sicard et al., 2020). Some of the other strains displayed a more extreme behavior. For example, Xfm CFBP 8173 exhibited a high growth and susceptibility to antimicrobial peptides, but low biofilm formation, which was not affected by NAC. In contrast, Xfp DD1 displayed a slow growth and formed abundant biofilm as observed in other studies (D'Attoma et al., 2020), but this biofilm was susceptible to NAC. Nevertheless, this strain is highly resistant to the tested antimicrobial peptides as it was previously reported (Baró et al., 2020b). Xfm IVIA 5901.2 exhibited a comparable growth pattern to that of Xfp DD1, but its susceptibility to antimicrobial peptides was similar to that of the other IVIA strains. Taking into account all these results, Xff IVIA 5387.2 was selected for next bactericidal and antibiofilm studies, because it presents a moderate behavior. Biofilm formation kinetics was assessed for this strain and maximum biofilm was formed between the 4th and 6th day in PD3 medium. This pattern was similar to the one previously reported for other X. fastidiosa strains (Cogan et al., 2013;Janissen et al., 2015). It corresponds to a typical biofilm formation kinetics, involving attachment of cells to a surface, EPS matrix secretion, biofilm formation, and biofilm maturation. Eventually, biofilm cells revert to a planktonic state and they are able to disperse.
The peptides tested in this study, including the reference peptide cecropin B, showed different degrees of bactericidal activity against X. fastidiosa, being classified into five major groups. The most interesting sequences were cecropin B, magainin 2, 1036, and RIJK2, which displayed similar activity with a reduction in viability higher than 3.2 log. Cecropin B and magainin 2 had been previously reported as active against X. fastidiosa (Li and Gray, 2003;Kuzina et al., 2006). However, this is the first report on the activity of 1036 and RIJK2 against X. fastidiosa. In fact, 1036 was previously reported to be active against P. aeruginosa and B. cenocepacia, and for RIJK2, only antibiofilm activity was described (De La Fuente-Núñez et al., 2012. It is interesting to highlight the difference in activity of RIJK2 compared to their analogs. For example, RIJK2 exhibited higher bactericidal activity than its all L-isomer RJK2, which could be ascribed to an increase in the stability of RIJK2 due to the presence of D-amino acids into its sequence as previously described for other peptides (Guell et al., 2011;Molhoek et al., 2011;Carmona et al., 2013). Moreover, this increased stability of RIJK2 could result in a reduction of its degradation susceptibility to the enzymes that X. fastidiosa secretes through outer membrane vesicles or through the type II secretion system (Ionescu et al., 2014;Rapicavoli et al., 2018;Feitosa-Junior et al., 2019). Nevertheless, more studies should be performed to confirm these observations. Most of the peptides tested in this work showed high antibiofilm activity against X. fastidiosa. Although some of them, such as RR4-OH, RIJK2, and 1036, had been previously described to display antibiofilm activity against Gram-negative bacteria, this is the first time that their activity against X. fastidiosa is reported (De La Fuente-Núñez et al., 2012Mohamed et al., 2017). Remarkably, we also identified peptides that had never been described to display antibiofilm activity. Among them, we found magainin 2 and indolicidin, only previously reported for their antibacterial activity against X. fastidiosa (Kuzina et al., 2006), and the newly designed peptides such as BP526 and RR4-NH 2 .
Similarly to other antibiofilm peptides against human pathogens (Mishra and Guangshun, 2017;Park et al., 2019;Qi et al., 2020), the antibiofilm activity of BP525, 1037, and R-FV7-I16 showed a dose-effect relationship that fitted well with a Michaelis-Menten saturation curve. Interestingly, they showed low ED 50 values, which means that low peptide concentrations already display high antibiofilm activity. This result suggests that, in a hypothetical plant application, the dilution of the peptides along the xylem vessels would not significantly affect their antibiofilm activity. Taking into account that these three peptides belong to different families of compounds, a similar behavior could be expected for the other peptides.
Interestingly, peptides that displayed high antibiofilm activity showed different patterns concerning the amount of planktonic cells detected during the screening of antibiofilm activity. This could indicate that the effect of these peptides on the biofilm formation may differ between them. The effect of BP527, 1026, and RJK2 in the biofilm formation was studied in detail. Peptides 1026 and RJK2 exhibited antibiofilm activity, because most of X. fastidiosa cells remained in a planktonic stage preventing biofilm formation. In the case of BP527, antibiofilm activity was also observed, but less planktonic cells were detected. This could suggest that this peptide displayed its activity once the biofilm was formed by causing a detachment of biofilm cells. Therefore, this could indicate that peptides are able to affect biofilm formation of X. fastidiosa at different stages whether by directly preventing biofilm formation or by affecting the reversible/irreversible attachment phase. This behavior has been reported for other pathogens such as P. aeruginosa and A. baumannii when treated with FLIP7 or ciprofloxacin (Macia et al., 2014;Gordya et al., 2017;Raheem and Straus, 2019). Nevertheless, further studies are needed to elucidate the exact role of these peptides in the inhibition of biofilm formation.
Regarding the hemolytic activity and the leaf infiltration effect of the peptides, it was not possible to establish a general pattern. In general, the peptides showed low hemolytic activity, and their effect upon infiltration on tobacco leaves was moderate and significantly lower than that of the reference peptide. It has to be taken into account that the effect observed in tobacco leaves might not necessarily be due to phytotoxicity, but it might be related to a hypersensitivity reaction caused by the peptides (Badosa et al., 2013). The least toxic families were those derived from 1036 and FV7. In the case of lipopeptides BP526-BP528, it is interesting to note that an increase of the fatty acid chain length led to an increase of the hemolysis. This correlation has been attributed to an increase of the peptide hydrophobicity that favors its affinity for the erythrocytes membrane (Malina and Shai, 2005;Oliveras et al., 2018).
To summarize, peptides with bactericidal and antibiofilm activity against X. fastidiosa and moderate toxicity have been identified. A PCA allowed to classify these peptides into four groups according to their distinct biological activity profile. An interesting group was composed by 1036, RIJK2, and magainin 2 as they displayed dual activity (high bactericidal and antibiofilm activities) and moderate toxicity. Another group with many peptides displayed high antibiofilm activity, but low/moderate bactericidal activity and a low toxicity profile. Peptides 1036 and RIJK2, with dual activity against X. fastidiosa and moderate toxicity, would be the most promising ones as they may be able to simultaneously inhibit biofilm formation and kill X. fastidiosa cells. Nevertheless, peptides with only antibiofilm activity should also be taken into account as they may be able to eliminate the symptoms caused by the occlusion of the xylem vessels by X. fastidiosa. However, this could cause an increase of planktonic cells available for vector transmission (Ionescu et al., 2014). Moreover, these peptides could be used in combination with other antimicrobial peptides in order to reduce the planktonic cells. Therefore, in future experiments, the most promising peptides identified in the present work will be tested in planta in different hosts to determine their capability to control the diseases caused by X. fastidiosa.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.