Complete killing was confirmed by plating an aliquot onto PW plates

Recent outbreaks of X. fastidiosa diseases in Europe and Asia and also in new plant hosts such as olive , blueberry , and pear suggest the great adaptation potential of this pathogen. In a number of plant species, X. fastidiosa is believed to live as a harmless endophyte without inducing disease symptoms . Coexistence in the same xylem system of different strains for a long time without killing the host represents a fertile environment for exchange of DNA material. Several MLST-based studies detected inter subspecific recombination among strains of X. fastidiosa and proposed recombination as the mechanism of new allele acquisition, leading to plant host shift and disease emergence. Intersubspecific recombination was described to generate strains that infect citrus and coffee , mulberry , and blueberry and blackberry . A recent study also showed intersubspecific recombination between coffee-infecting strains in plants intercepted in France . Natural competence could be an explanation for the frequent recombination events detected in X. fastidiosa. Natural competence in X. fastidiosa was recently described in vitro , the rate of homologous recombination was shown to be higher when the cells were growing exponentially in solid agar plates than in batch culture tubes, and minimal medium was more conducive than rich medium .

With a plasmid as a donor DNA, 96 bp of flanking homology was sufficient to initiate recombination . Moreover,vertical farm tower some competence-related and type IV pili genes were shown to be involved in the process . Although some of those studies were performed using plasmids as donor DNA, two strains were also shown to recombine in coculture conditions , although the capacity of these strains to act either as a donor or a recipient for DNA exchange was not determined in those studies. The objective of this study was to test the hypothesis that natural competence in X. fastidiosa occurs under flow conditions . Associated with this objective was the aim of elucidating whether previous observations of high frequencies of X. fastidiosa natural competence in vitrowere dependent on batch culture conditions , which allow cell-tocell contact for longer times without replenishing of nutrients or removal of secreted molecules. Although natural competence and recombination are assumed to occur in natural habitats based on field surveys and DNA sequence data, experimental indications of its occurrence in the plant or insect host are not yet available for X. fastidiosa. Therefore, to circumvent the limitation of X. fastidiosa recombination tests in the natural hosts that are affected by uneven bacterial distribution and low populations , we performed recombination experiments in a microfluidic chamber system that mimics the natural environment of xylem vessels and insect foreguts.

The MC system allows continuous media flow conditions and formation of biofilms and has been previously used to study the behavior of X. fastidiosa . The biofilm fraction of the MC and the planktonic and detached cell fraction can be collected separately, and the behavior of cells in the two fractions can be determined. Two strains used in all of the previous publications on this topic were used in this current study to facilitate comparison with the literature and to further our understanding of natural competence in X. fastidiosa. The results presented here show that growth under flow conditions supports natural competence in X. fastidiosa, with recombination frequencies equivalent to that on solid media, previously described to be the most conducive environment for natural competence in vitro . These findings support the hypothesis that recombination occurs at high rates under flow conditions, representing the natural habitats of X. fastidiosa.Xylella fastidiosa subspecies fastidiosa mutants NS1-CmR and pglA-KmR were used in this study. The mutants were cultured in periwinkle wilt agar medium , modified by omitting phenol red and adding 1.8 g liter 1 bovine serum albumin and supplemented with the respective antibiotics. PD3 medium and modified X. fastidiosa medium were used when stated. Pectin was added to a final concentration of 0.01% as previously described . Kanamycin was used at 30 g m1 and chloramphenicol at 10 g ml. Inocula were prepared by streaking cultures from the 80°C freezer stocks on PW agar plates and incubating the plates for 5 to 7 days at 28°C. Cultures were then rest reaked onto new plates and incubated for another 5 to 7 days before use.To select a medium to test the occurrence of natural competence in MCs, three media were first tested in solid agar plates.

XFM and PW, used in previous studies , were selected as positive- and negative-control media, respectively, for recombination. Natural competence experiments were performed according to the method of Kung et al. with some modifications. Briefly, cells of the NS1-CmR and pglAKmR mutants were prepared in liquid media by scraping the cultures from PW-antibiotic plates. Ten microliters of each strain was spotted on top of each other on the agar plates of PW, XFM, and PD3 without antibiotics, and the spots were allowed to dry for 1 h. The plates were then incubated at 28°C for 3 days. Next, two spots from the same plate were scraped off and suspended in 1 ml of PD3 to make one replication, and 3 to 4 replicates were included for each media type per experiment. The experiments were repeated independently twice for XFM and at least three times for PD3 and PW. Single mutant strains were included as controls. The suspensions were then serially diluted, and100 l aliquots of appropriate dilutions were plated on PW agar plates in triplicate supplemented with both antibiotics to recover recombinants at the antibiotic-resistant site and with a single antibiotic to check for the growth of both parents in the mixture. Appropriate dilutions also were plated onto PW plates without antibiotics for enumeration of total viable cells. Plates were incubated at 28°C for at least 14 days before CFU were enumerated. The recombination frequency at the antibiotic-resistant site was calculated as the ratio of recombinant CFU to total CFU in equal volumes of suspension. After selection of the media that supported recombination in the agar plates, the media were tested in the MCs for cell attachment and biofilm formation.To test for the specific components that may influence natural competence, an initial screen was performed by removing or adding components to PW and PD3 in solid agar plates as described above. The components tested were sodium citrate dehydrate, succinic acid, and starch ; BSA and L-glutamine ; and pectin.

The effect of BSA was further tested by supplementing PD3 and PD3 plus L-glutamine with BSA and removing BSA from PW and XFM. Experiments were repeated three times independently with three replicates each time, except for PD3 plus L-glutamine treatment that was performed once with three replicates. The twitching motilities of both mutants were determined in media with and without BSA, according to previous studies with few modifications. Briefly, for PD3 and PW with and without BSA, media plates solidified with agar or Gelrite were divided into two halves, 10 to 12 spots of each mutant strain were made using a sterile toothpick, and plates were incubated at 28°C for 4 to 5 days. For XFM with and without BSA, plates solidified with agar were used and incubated for 10 to 12 days before measurements were recorded. Colony peripheral fringes were observed under 10 magnification using a Nikon Eclipse Ti inverted microscope ,vertical planter tower and fringe widths were measured for six colonies per plate per strain, with at least seven measurements per colony using a Nikon DS-Q1 digital camera connected to a Nikon Eclipse Ti inverted microscope and controlled by NIS-Elements imaging software version 3.0. Twitching experiments were performed at least three times independently for PD3 and PW with and without BSA and once for XFM with and without BSA.Three growth conditions were used: solid agar plates , liquid culture tubes , and continuous liquid flow . PD3 without antibiotics was the medium used, and the initial inocula of the NS1-CmR and pglA-KmR mutants were prepared as described above. Competence in tubes. Twenty-five-milliliter glass test tubes containing 3 ml of PD3 were inoculated with 100 l of each of the OD adjusted strain suspensions as donor and recipient cells . Tubes containing single strain inoculations were included as control treatments. Tubes were then incubated with shaking . After 3 days, the tubes were vortexed well to mix the biofilm formed on the air-liquid interface with the rest of the suspension and serially diluted and plated as described above. Three independent experiments were performed, and three replications were included in each experiment .MCs were prepared as previously described . Briefly, two parallel channels with separate inlets for bacterial cells and growing media were etched on a silicon wafer. The channels were modeled into polydimethylsiloxane and sandwiched between the PDMS layer and a glass cover slide. The inlets and outlets were then connected to tubings that were connected to syringes . The syringes were connected to pumps which control the media flow rate in the MC. The MC was mounted onto a Nikon Eclipse Ti inverted microscope to observe cell attachment and biofilm formation using phase-contrast and Nomarski differential interference contrast optics. Time-lapse video was taken using a Nikon DS-Q1 digital camera connected to the microscope and controlled by NIS-Elements imaging software version 3.0. For preparing the inocula for MCs, equal volumes of the strain pairs were mixed and inoculated into the cell inlet syringes, and growing medium was injected in the media syringes. MCs were run for 5 to 7 days with a media flow rate of 0.25 l min1 until abundant growth of biofilm was observed.

At the end of the experiment, the fraction of cells collected in the outlet syringe was harvested, and the fraction formed inside the channels was detached and pushed to the outlet collection syringe by increasing the flow rate to 30 to 40 l min 1 . Serial dilution, plating, CFU counts, and the frequency of recombination calculations were done as described above. Four independent experiments were performed with seven replicates in total .Confirmation of homologous recombination occurring via natural competence was performed by using heat killed donor cells in the solid agar plates. Suspensions of the donor cells were incubated at 90°C for 15 min for heat killing. The heat-killed donor and live recipients were then spotted on PD3 plates as described above. For confirmation of homologous recombination at the desired genome region, randomly selected recombinant CFU were rest reaked onto new double-antibiotic PW plates, and colony PCR was performed using the primers targeting the flanking region of the construct used to generate the mutants according to Kung et al. . Sequences of the flanking regions of antibiotic cassette insertion sites between the parent strains were compared using the muscle pairwise alignment algorithm within the Geneious 9.0.3 platform .Several hypotheses have been proposed to explain the existence of natural competence in bacteria. One explanation is that starvation signals induce competence, and the incoming DNA serves as a nutrient source under poor nutrient conditions as demonstrated in H. influenzae , Pseudomonas stutzeri , and R. solanacearum . Based on the results with a minimal medium and a rich undefined medium , a previous study speculated that growth in a low-nutrient medium favors natural competence inX. fastidiosa. However, the results of this study with these two media and PD3, another undefined rich medium, demonstrated that growth in PD3 significantly increases the recombination frequency. This suggests that starvation is not necessary to induce competence in X. fastidiosa. Further investigations of the differences between PD3 and PW were performed by either removing or adding these components to/from one another. Initial screening with the components showed a pronounced effect of BSA on the number of recombinants recovered. Additional experiments confirmed that BSA significantly reduces the recombination frequency when present in PD3, PW, and XFM. Since both XFM and PW contain BSA, this may explain the lower recombination frequencies in these media. In a previous study, BSA had been found to reduce the surface attachment and twitching motility of X. fastidiosa . In fact, natural competence and twitching motility are dependent on the activity of type IV pili in X. fastidiosa . Therefore, in this study the correlation between twitching movement and natural competence in different media was investigated. Interestingly, PD3 allowed the highest fringe width, and the presence of BSA significantly reduced twitching motility in all three media.