Natural Competence and Recombination in the Plant Pathogen Xylella fastidiosa

Abstract

Homologous recombination is one of many forces contributing to the diversity, adaptation, and emergence of pathogens. For naturally competent bacteria, transformation is one possible route for the acquisition of novel genetic material. This study demonstrates that Xylella fastidiosa, a generalist bacterial plant pathogen responsible for many emerging plant diseases, is naturally competent and able to homologously recombine exogenous DNA into its genome. Several factors that affect transformation and recombination efficiencies, such as nutrient availability, growth stage, and methylation of transforming DNA, were identified. Recombination was observed in at least one out of every 10⁶ cells when exogenous plasmid DNA was supplied and one out of every 10⁷ cells when different strains were grown together in vitro. Based on previous genomic studies and experimental data presented here, there is mounting evidence that recombination can occur at relatively high rates and could play a large role in shaping the genetic diversity of X. fastidiosa.

INTRODUCTION

Naturally competent bacteria, which are able to uptake DNA under natural growth conditions, are found in a wide range of phyla, suggesting that this trait is functionally important and could confer a fitness benefit (41). For these bacteria, DNA acquired through natural transformation could recombine into the genome, providing a source of genetic diversity and potentially a means of horizontal gene transfer, although it is possible that natural transformation evolved as a nutrient uptake system (36). Many factors affect the onset of competence in different bacteria. Bacteria can become competent in response to environmental signals or cues, such as antibiotics or alkaline conditions (10). Nutritional factors can also play a role. For example, the presence of chitin induces competence in Vibrio cholerae (31), while starvation conditions induce competence in Haemophilus influenzae (27). Growth stage can also be a regulating factor (10). A notable outlier, however, is Neisseria gonorrhoeae, which is not known to regulate its competence and is able to acquire DNA during all phases of growth (21). Recently, interest has risen in the consequences of horizontally transferred DNA on the evolution of microbial pathogens (2).

Populations of plant pathogenic bacteria, especially those colonizing crops, are faced with unique environmental pressures that impact the genetic diversity observed for populations. Because host plants tend to be genetically similar, pathogens may undergo periodic selective sweeps in which allelic diversity is reduced or eliminated by the emergence of a genotype with increased fitness (18). In practice, such a process mimics the effects of bottlenecks on genetic diversity, as clonal individuals with a shared common ancestor dominate the population. Host-specialized plant pathogens may undergo selective sweeps, eventually embarking on a coevolutionary arms race with host plants (11). However, a different scenario is plausible for generalist plant pathogens. For these organisms, homologous recombination may lead to the generation and maintenance of allelic diversity in populations (22), which would potentially permit them to explore a wider variety of host plants. While natural competence has been observed for the generalist plant pathogen Ralstonia solanacearum (4), it has not been documented for other bacterial plant pathogens. However, there is mounting evidence that recombination of homologous DNA acquired through natural transformation and other mechanisms could play a much larger evolutionary role than previously thought. Recombination has been shown to increase the rate of pathogen adaptation (3), and emerging diseases have been attributed to the horizontal transfer of virulence factors (19).

Xylella fastidiosa is a plant-pathogenic bacterium that colonizes the xylem vessels of a wide range of host plants and the foregut of its leafhopper vectors (6). Pathogenicity apparently results when X. fastidiosa reaches a high population density and moves between xylem vessels, inhibiting the flow of xylem sap and leading to symptoms such as leaf scorching and stunted growth (30). X. fastidiosa colonization of plants is strain and host species dependent. In some cases, the bacterium multiplies, moves systemically, and induces disease symptoms. In other hosts, X. fastidiosa multiplies somewhat but does not move systemically, while in yet another group of host plants, the bacterium may move systemically but not cause disease (35). X. fastidiosa is the causative agent of Pierce's disease in grapevines in addition to several emerging diseases in other plant hosts, such as citrus variegated chlorosis and coffee and oleander leaf scorch (23). Recent studies have shown evidence of recombination between different strains of X. fastidiosa (1, 39, 45). In addition, multilocus sequencing typing (MLST) studies have indicated that horizontally acquired sequences may play a significant role in introducing genetic diversity to X. fastidiosa populations, potentially being more important than point mutations (1, 39). Scally et al. (39) estimated the ratio of the contribution toward diversity of recombination and point mutations (r/m) in X. fastidiosa to be 3.23 at the nucleotide level.

Based on increasing evidence that recombination significantly contributes to the genetic diversity of X. fastidiosa, we hypothesized that this organism can acquire DNA through natural transformation. This study documents natural competence and the homologous recombination of acquired DNA in X. fastidiosa and reports several factors that affect its occurrence.

MATERIALS AND METHODS

Strains, DNA, media, and growth conditions.

X. fastidiosa subspecies fastidiosa strains Temecula (42) and STL (16) were used in this study. The knockout mutant of the pglA gene (in strain Fetzer), which encodes a polygalacturonase necessary for systemic colonization of grapevines, was obtained from Roper et al. (37); the knockout mutant of the rpfF gene, which encodes the synthase for the cell-cell signaling fatty acid molecule diffusible signaling factor (DSF), was obtained from Newman et al. (32); the knockout mutant of the rpfC gene, which encodes a two-component regulatory protein that senses DSF, was obtained from Chatterjee et al. (7). The above mutants are all kanamycin resistant. Strain NS1-CmR, which contains a chloramphenicol resistance cassette inserted in a noncoding region of the genome, was obtained by transforming Temecula with pAX1-Cm from Matsumoto et al. (29). DNA used for transformation experiments consisted of suicide plasmids pAX1-Cm (29) and pKLN61 (32) or an approximately 2-kb linear segment of pKLN61 containing the kanamycin resistance cassette PCR amplified using the forward and reverse primers rpfF-fwd and rpfF-rev, respectively (Table 1 ). All strains were grown in periwinkle wilt Gelrite (PWG) medium or periwinkle wilt (PW) medium (liquid medium without solidifying agent) (13) or in X. fastidiosa medium (XFM) (24) with the following modifications: 1 g/liter K₂HPO₄ and 0 g/liter KH₂PO₄. Preliminary data suggested that transformation with pKLN61 is more efficient in modified XFM. When appropriate, kanamycin was added to a final concentration of 30 μg/ml and chloramphenicol was added to a final concentration of 10 μg/ml.

Table 1.

Primers used in this study

Primer function	Primer namea	Sequence (5′ to 3′)	Source or reference
Flanking rpfF gene	rpfF-fwd	TGGAGTGGTGTGCTCTTGTCCA	This study
	rpfF-rev	ACGCGATACGGAAGTACCACCA
Flanking rpfC gene	rpfC-fwd	AGCTTTTGGTGTTGCTGTCCG	This study
	rpfC-rev	GCTTTCATCGTAAAACCCCACTG
Flanking pglA gene	pglA-fwd	AATCGCTCAGCTTCAGTCCG	This study
	pglA-rev	GCCCTCCAGTGAAGGAATTTCT
Flanking NS1	NS1-f	GTCAGCAGTTGCGTCAGATG	30
	NS1-r	AAAGCTGCCGACGCCAAATC
X. fastidiosa quantification	HL5	AAGGCAATAAACGCGCACTA	17
	HL6	GGTTTTGCTGACTGGCAACA

^afwd or f, forward; rev or r, reverse.

General transformation protocol.

X. fastidiosa cells were harvested from solid PWG medium after approximately 7 days of growth and diluted in 200 μl modified XFM to a final optical density at 600 nm (OD₆₀₀) of between 0.0025 and 0.05 (approximately 10⁶ to 2 × 10⁷ CFU/ml). After 2 days of growth at 28°C with constant shaking at 180 rpm, DNA was added to a final concentration of 5 μg/ml. Cultures were then grown for an additional 24 h and plated on selective media. Antibiotic-resistant colonies were counted after approximately 14 days of growth, and recombination events were confirmed through PCR analysis of random samples. Primers rpfF-fwd and rpfF-rev (Table 1) were used to confirm the presence of the kanamycin resistance cassette in the rpfF region of the genome, and primers NS1-f and NS1-r (29) were used to confirm recombination between pAX1-Cm and the NS1 region of the genome.

DSF detection assay.

Detection of DSF in the X. fastidiosa growth medium was conducted using the Xanthomonas campestris green fluorescent protein (GFP)-based reporter Xcc 8523 (pKLN55) (32), which expresses GFP in the presence of DSF. DSF was extracted from 10 PWG plates each grown for 1 week with wild-type Temecula, the rpfF mutant, or recovered antibiotic-resistant cells transformed with either pKLN61 or the linear, PCR-amplified fragment. DSF was extracted essentially as described by Newman et al. (33). Plates were homogenized with 100 ml of water-saturated ethyl acetate and incubated for 1 h at room temperature. The supernatant was collected and evaporated, and DSF was recovered from the ethyl acetate precipitate into 1 ml of methanol. DSF extract (50 μl) from each sample was spotted on a paper disk on a King's B agar (KB) plate, and 100-μl aliquots of the Xcc 8523 (pKLN55) were streaked perpendicular to the disk. The plates were incubated overnight at 28°C, and fluorescence was viewed using a Zeiss SV11 stereoscope with Kramer epifluorescence/Optronix Color DEI450.

Cell density recombination experiments.

To determine the effect of cell culture age on recombination, X. fastidiosa cells were diluted to an OD₆₀₀ of approximately 0.007 in modified XFM and PW medium. pKLN61 was added to a final concentration of 5 μg/ml after 2, 4, 6, or 8 days of growth. Cultures were grown for an additional 24 h, and 50-μl aliquots were plated on PWG medium and on PWG medium with kanamycin. Antibiotic-resistant colonies were counted approximately 14 days after plating. To quantify growth of X. fastidiosa in modified XFM, cultures with an initial OD₆₀₀ of 0.005 were grown in triplicate at 28°C with constant shaking at 180 rpm. Samples were taken at days 0, 1, 2, 3, 5, 7, and 9, and the total cell count was estimated using quantitative PCR as described below. To determine the effect of initial cell density on recombination efficiency, cultures were prepared as described above, with initial OD₆₀₀s of 0.0025, 0.005, 0.01, 0.025, and 0.05. pKLN61 was added after 2 days of growth, and cells were plated on PWG medium with kanamycin after an additional 24 h. Samples were also taken for quantitative PCR to determine the total number of cells present. Six replicates were used for the recombination experiments.

Recombination between X. fastidiosa strains in vitro.

The rpfF, rpfC, and pglA mutants, which are kanamycin resistant, were each individually grown in triplicate with strain NS1-CmR, which is chloramphenicol resistant, in test tubes with 3 ml of modified XFM. The mutant strains and NS1-CmR were each diluted to an OD₆₀₀ of 0.005 in the same tube, for a total initial OD₆₀₀ of 0.01. Cultures (initial OD₆₀₀, 0.01) of rpfC, rpfF, and pglA mutants and NS1-CmR were also grown individually as controls. Cultures were grown for 3 days at 28°C with constant shaking at 180 rpm. Fifty-microliter aliquots from days 0 and 3 were plated on PWG medium with kanamycin, PWG medium with chloramphenicol, and PWG medium with kanamycin and chloramphenicol. Population sizes at each time point were estimated with quantitative PCR (see below). Colonies that were resistant to both kanamycin and chloramphenicol were PCR screened for the presence of the antibiotic resistance cassettes in the appropriate loci using primer pairs pglA-fwd and -rev, rpfF-fwd and -rev, rpfC-fwd and -rev (Table 1), and NS1-f and NS1-r (29).

X. fastidiosa cells and extracellular DNA quantification.

Quantification of X. fastidiosa cells and extracellular DNA was performed using quantitative PCR. For cell quantification, aliquots of culture were heated at 99°C for 15 min, and 1 μl of the suspension was used as the template for quantitative PCR using Applied Biosystem's Sybr green master mix with 0.5 μM (each) primers HL5 and HL6 (17). Reactions (absolute quantification) were run on an Applied Biosystems 7500 Fast in standard mode with the following cycles: 55°C for 2 min, 95°C for 10 min, and 40 cycles of 95°C for 10 s, 60°C for 30 s, and 72°C for 30 s. Fluorescence was quantified at 60°C. Standards were created using cells harvested from PWG medium and diluted in modified XFM. Cell counts for standards were determined based on OD₆₀₀. To quantify extracellular DNA, the culture was passed through a 0.2-μm filter to remove cells, and 1 μl of filtrate was used as the template. Standards were created with genomic Temecula DNA quantified using a Thermo Scientific Nanodrop 1000 Spectrophotometer.

Transformation efficiency of methylated plasmids.

Previous work has shown that X. fastidiosa has a functional restriction modification system, and properly methylated plasmids recombine at higher efficiencies when introduced through electroporation (28). We transformed Escherichia coli strain EAM1 (28), which expresses the X. fastidiosa methylase PD1607, with pKLN61 using a Z-Competent E. coli transformation buffer set from Zymo Research (Irvine, CA). Cells were then grown overnight in LB with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) to induce expression of the methylase. Transformation experiments were conducted with a starting OD₆₀₀ of 0.01.

RESULTS

X. fastidiosa is naturally competent.

X. fastidiosa strain Temecula was able to naturally transform and homologously recombine exogenous DNA into its genome when grown in modified liquid XFM without selection. After adding pKLN61, a plasmid that cannot replicate in X. fastidiosa containing portions of the rpfF gene flanking a kanamycin resistance cassette, or a linear PCR-amplified copy of the gene and cassette, we recovered antibiotic-resistant colonies. PCR analysis confirmed that the antibiotic-resistant colonies were a result of a double recombination event between the genome and plasmid/linear DNA (Fig. 1A). Amplification of the rpfF locus of the transformed colonies and the rpfF mutant (control) resulted in a single amplicon approximately 800 bp longer than the wild-type copy. Complete inactivation of the rpfF gene was also confirmed using Xanthomonas campestris strain Xcc 8523 (pKLN55) that expresses GFP in the presence of DSF (Fig. 1B to D).

Fig. 1.

Confirmation of natural competence and recombination in X. fastidiosa. (A) Size standards (lane 1) from top to bottom: 2, 1.6, and 1 kb. PCR of a negative control (lane 2) and genomic DNA extracted from colonies transformed with pKLN61 (lanes 3 and 4), linearized rpfF:kanR (5 and 6), the rpfF mutant (7), and wild-type Temecula (8) was performed using primers rpfF-fwd and rpfF-rev. The wild-type copy of the rpfF gene is approximately 1.2 kb, and the mutant copy is approximately 2 kb. Fluorescence microscopy images of bioassay detection for the presence of DSF in extracts taken from wild-type cells (B), a colony transformed with pKLN61 (C), and the rpfF mutant (D). GFP expression by Xcc 8523 indicates the presence of DSF, which is not produced by the strain transformed with pKLN61 or the rpfF mutant.

The transformation protocol was also performed using strain STL and different plasmids in Temecula to show that the phenomenon of natural transformation and recombination is not strain or sequence specific. STL was successfully transformed with pKLN61 and the linear PCR-amplified fragment. Temecula was also successfully transformed with pAX1-Cm (29), which inserts a chloramphenicol resistance marker in a noncoding region of the genome, and several other novel plasmids. PCR analysis confirmed clean double recombination events in all cases (data not shown).

Cell growth affects recombination efficiency.

The maximum number of recombinants, as determined by acquired antibiotic resistance, was obtained when DNA was added after 2 days of growth on modified XFM. Transformation and recombination rates decreased with time after inoculation. Cells were grown for 2, 4, 6, or 8 days before pKLN61 was added. No antibiotic-resistant colonies were recovered when DNA was added after 8 days of growth (Fig. 2A). Cells were also grown in PW medium with pKLN61, but no recombinants were recovered at any time point. In addition, no antibiotic-resistant colonies were recovered for the control group when no DNA was added. However, all treatments grew on PWG medium without antibiotics, confirming that cell cultures were still viable. A growth curve for cells in XFM shows that there is an initial 2-day lag phase under experimental conditions, and the cells enter stationary phase within 7 days after inoculation (Fig. 2B). Cells appear most competent as they enter exponential growth, as the number of recombinants recovered drops dramatically when DNA is added after 4 and 6 days, despite the fact that more cells are present.

Fig. 2.

Growth effects on recombination efficiency. (A) The number of recombinants decreased over time as pKLN61 was added after 2, 4, 6, or 8 days of growth. Cultures were plated on selective media after a 24-h incubation period with transforming DNA. Recombination events were quantified based on the number of antibiotic-resistant colonies present after approximately 14 days of growth. (B) Growth curve of cells in XFM. (C) Total number of recombinants recovered from cultures with different starting optical densities. Cultures were grown for 2 days before the addition of pKLN61 and grown for an additional 24 h before plating on selective media. (D) Recombination efficiency (squares, dashed line) and number of generations over 3 days (triangles, solid line) for cultures with different starting ODs. The correlation coefficient (r) between the log-transformed recombination efficiencies and number of generations is 0.826 (P < 0.001). Different symbols/letters indicate statistically significant differences among treatments (P < 0.05).

To determine if cell density was solely responsible for the decrease in antibiotic-resistant colonies recovered over time, the recombination efficiency (number of recombinants recovered divided by the final cell count) for cell cultures with different initial optical densities was measured. Cell cultures with initial OD₆₀₀s from 0.0025 to 0.05 were given pKLN61. No recombinants were recovered from cultures with an initial optical density of 0.0025 (Fig. 2C). For cultures with higher starting optical densities, however, there was an increase in the total number of recombinants recovered per 50 μl. An analysis of variance and Tukey statistical test run on the log-transformed recombination efficiencies indicates that the efficiency for an OD₆₀₀ of 0.0025 was significantly lower than those for all other optical densities (P = 2.2 × 10⁻¹⁶), but the difference in recombination efficiencies between the other treatments was not significant (Fig. 2D). Overall recombination efficiencies were on the order of 4 to 10 recombinants for every 10⁶ cells present.

In a comparison of the starting and final cell counts for each of the cultures, the data show that the number of generations differed depending on the starting optical density. For cultures with a starting OD₆₀₀ of 0.0025, there was an average of 1.3 generations over the course of 3 days (Fig. 2D). The average number of generations peaked at 3.4 for cultures with an initial optical density of 0.01. Growth appeared to be slightly inhibited when cells were started at the highest OD₆₀₀ of 0.05, as the total number of generations over 3 days decreased to approximately 2.4. Cultures that started with an OD₆₀₀ of 0.0025 had significantly fewer generations over the course of 3 days than all other treatments. Cultures that started with an OD₆₀₀ of 0.05 also underwent significantly fewer generations than cultures with an initial OD₆₀₀ of 0.01. There was no significant difference in any of the other treatments. The correlation coefficient (r) between the log-transformed recombination efficiency and number of generations was 0.826, with a P value of <0.001 (28 degrees of freedom).

Recombination between X. fastidiosa strains in vitro.

Recombination occurred between strains of X. fastidiosa when different strains were cocultured. X. fastidiosa knockout mutants of rpfF, rpfC, and pglA (all are kanamycin resistant) were individually inoculated with strain NS1-CmR (chloramphenicol resistant) in liquid XFM and grown for 3 days at 28°C. Recombination between the mutant and NS1-CmR was quantified by counting the number of doubly antibiotic-resistant colonies (Fig. 3). PCR analysis confirmed site-specific recombination by the presence of a single, antibiotic-resistant allele at each of the appropriate loci. Recombination was observed in approximately 4 out of every 10⁷ cells for the pglA mutant and NS1-CmR, and 9 out of every 10⁷ cells for the rpfC mutant and NS1-CmR. No recombination was detected between the rpfF mutant and strain NS1-CmR. An analysis of variance test indicated that the three treatments were significantly different (P = 0.018), but a multiple comparisons of means (Tukey) indicated that the difference in recombination efficiencies between mutants rpfC and pglA with NS1-CmR was not significant (P = 0.13).

Fig. 3.

Recombination efficiencies between different strains grown in coculture. Mutant pglA, rpfF, and rpfC strains were grown individually in coculture with NS1-CmR in liquid XFM for 3 days. Recombination efficiencies between strains are shown based on the number of colonies resistant to both kanamycin and chloramphenicol per total number of cells present (solid bars) and nanograms of extracellular DNA in the medium (empty bars). Recombination events were also confirmed by PCR analysis. Different letters indicate statistically significant differences among treatments (P < 0.05).

Extracellular DNA present in the media at the start of the experiment and after 3 days of growth was also quantified. Aliquots of each culture at 0 and 3 days were filtered through a 0.2-μm filter to remove cells, and total extracellular DNA was estimated using quantitative PCR. The amount of extracellular DNA increased after 3 days for all three treatments—approximately 15, 36, and 37 times for the rpfC, rpfF, and pglA mutants grown with NS1-CmR, respectively. There was no significant difference in the amounts of extracellular DNA present in the different treatments after 3 days of growth (P = 0.099), with the final concentration being 6 to 22 picograms per microliter. In addition, there was no significant difference in the recombination efficiency when measured in terms of extracellular DNA concentration (P = 0.170) (Fig. 3).

Transformation and recombination of methylated DNA is more efficient.

As previous work has indicated that X. fastidiosa has a functional restriction modification system (28), the recombination efficiencies of cells transformed with pKLN61 and X. fastidiosa-specific methylated pKLN61 were compared. Significantly more recombinants were recovered from cells transformed with the methylated plasmid than with the unmethylated plasmid (average of 28.3 ± 5.15 and 6.3 ± 1.84 per 50 μl of culture, respectively [P = 0.003]).

DISCUSSION

X. fastidiosa appears to exchange DNA at relatively high rates when different strains are grown together in vitro, and actual recombination rates between strains were likely higher than observed. High recombination rates between cocultured strains have also been observed for other bacteria; recombination in mixed cultures of Streptococcus mutans and Streptococcus gordonii was observed for approximately one out of every 10⁶ viable cells after 4 h of incubation (25). As recombination between the X. fastidiosa strains occurred before the cells were subjected to selective pressure on the markers, and assuming recombination occurs randomly between homologous regions of the genome, one can postulate that only somewhere on the order of 0.1% of total events (based on X. fastidiosa's approximately 2.5-Mb genome and the 1-kb antibiotic marker) was detected. Previous work with Helicobacter pylori showed that recombination of naturally transformed DNA in approximately one out of every 10⁷ cells per passage (about 5.5 generations) was sufficient to confer a fitness advantage over noncompetent cells (3). Data indicate that natural transformation and recombination rates in X. fastidiosa are at least that high, suggesting that recombination can alter X. fastidiosa's genome and increase fitness and adaptation on a larger and more rapid scale than mutations alone.

Growth stage has been shown to be one of many factors regulating competence in bacteria, although there does not seem to be a single trend. In Streptococcus pneumoniae, for example, competence is inhibited during stationary phase, while in Bacillus subtilis, competence does not develop until its onset (10). Based on growth curve data, X. fastidiosa appears to be most competent as it is entering exponential growth, similar to R. solanacearum, for which the onset of competence occurs near the beginning of exponential growth and quickly declines during log phase; cells are essentially not competent once they reach stationary phase (4). This is also confirmed by the significant correlation found between the log-transformed transformation efficiencies and generations. The correlation suggests that the increase in the number of recombinants recovered from cultures that had undergone more generations was due at least in part to previously transformed cells multiplying (as opposed to new cells being transformed). Testing the effect of different starting optical densities on the transformation and recombination rates revealed that cell density does not seem to directly affect transformation rates on the time scale tested but indicates that cells must be dividing in order for recombination to occur. It is possible that no recombination and little growth for cultures with a starting OD₆₀₀ of 0.0025 was observed because there was an extended lag phase due to the low cell density. However, cell-cell signaling may still be important, as recombination was not detectable between the rpfF mutant and NS1-CmR when the two strains were cocultured in modified XFM. Quorum sensing has been shown to control competence in other bacteria (43), so it is possible that it at least partially regulates this process in X. fastidiosa.

In addition, we found that X. fastidiosa is competent when grown in a nutrient-limited, defined growth medium (modified XFM) but cannot transform and recombine exogenous DNA when grown in an undefined, rich medium (PW medium). Competence in other bacteria is often controlled by factors such as growth stage and nutritional signals (8). This seems to be the case for X. fastidiosa as well. The fact that X. fastidiosa appears to be most competent when it is undergoing rapid growth in low-nutrient conditions lends support to the hypothesis that natural transformation may be the result of cells importing nucleotides for nutritional purposes (36).

Data also indicate that the extracellular DNA concentration increases in cultures of X. fastidiosa. There are several possible reasons and mechanisms for this increase. It has been suggested that extracellular DNA plays a role in X. fastidiosa biofilm formation (9). Studies of S. pneumoniae have shown that when cells become competent, they can induce the lysis of noncompetent cells in the population to provide a source of nucleotides (10). Other naturally competent bacteria, such as N. gonorrhoeae, can donate DNA for transformation through type IV secretion systems (21). However, it is not known if the extracellular DNA in the case of X. fastidiosa is the result of dead cells with compromised membranes, competent cells actively inducing the lysis of other cells, or live cells secreting DNA. If one assumes that all actively growing X. fastidiosa cells increase the concentration of extracellular DNA in their immediate environment, recombination between different strains could happen readily in natural environments, such as in plants or insect vectors, whenever two different strains come in contact with each other. There is evidence that mixed infections can exist in X. fastidiosa's sharpshooter vectors (12), providing a possible means and location for recombination to occur in natural environments. MLST studies showing that recombination occurs between different strains of X. fastidiosa support this hypothesis (1, 39, 45). The fact that extracellular DNA concentrations probably increase during growth could affect our total cell quantification, as it was assumed that the DNA detected by quantitative PCR was from intact cells. However, if the number of cells present has been overestimated, then recombination rates would be even higher than reported.

Recombination rates were about five times higher when cells were transformed with X. fastidiosa-specific methylated plasmids than when cells were transformed with unmethylated DNA. Similar increases in recombination rates were found when methylated and unmethylated plasmids were introduced into X. fastidiosa by electroporation (28). Restriction modification plays a large role in DNA processing in many naturally competent bacteria; N. gonorrhoeae encodes approximately 16 methyltransferases in its genome and, with the corresponding endonucleases, acts as a restriction barrier to transformation of foreign plasmids. There is no similar restriction on genomic DNA, and transformation efficiencies can be 1,000-fold higher than those for foreign DNA (21). There is also evidence that the presence of different restriction modification systems coincides with the structure of phylogenetic clades in Neisseria meningitidis and could account for the differential barrier to DNA exchange observed within the species (5). Restriction modification systems could help explain how such high recombination rates were observed when X. fastidiosa strains were grown in coculture, despite the fact that extracellular DNA concentrations were low. Nonmethylated or improperly methylated DNA, such as what was used in initial experiments, may be subject to X. fastidiosa's restriction modification system and degraded before it can be recombined into the genome. Transformation experiments had approximately 1,000-fold more DNA added than was found in the coculture media. In addition, if the extracellular DNA released in coculture is simply genomic DNA, then less than one-thousandth of the potentially transforming DNA contained the selectable marker. Thus, it appears that X. fastidiosa cells are able to take up and recombine DNA originating from X. fastidiosa at much higher rates than they can for DNA from other sources. The active restriction modification system and the fact that X. fastidiosa has historically been grown in nutrient-rich media could help explain why this bacterium has been difficult to transform using conventional methods.

The experiments performed here provide evidence that DNA can enter X. fastidiosa cells through natural transformation and be homologously recombined into the genome. None of the DNA sources used could independently replicate in X. fastidiosa, so the antibiotic markers could not have been conjugally transferred between cells on a plasmid. Recent work has suggested that certain strains of X. fastidiosa harbor plasmids that encode conjugal transfer proteins and could potentially be horizontally transferred through conjugation (40), but in our case, the mechanism of horizontal gene transfer is transformation. A search of the X. fastidiosa genome found many genes and putative genes that would allow for the uptake and recombination of exogenous DNA (Table 2). Based on the mechanism of transformation in N. gonorrhoeae (21), a naturally competent Gram-negative bacterium, we propose that transforming DNA crosses the outer membrane of X. fastidiosa through a PilQ or PilQ-like channel that is part of a type IV or similar pilus. A type IV pilus or a pilus-like structure is required for DNA uptake in nearly all naturally transformable bacteria, as are a series of Com proteins (8). Recent work has shown that X. fastidiosa has functional type IV pili (14, 26), and a search of the X. fastidiosa genome produced several annotated com genes (Xylella fastidiosa Temecula1 genome page; http://cmr.jcvi.org/cgi-bin/CMR/GenomePage.cgi?org=ntxf02). The different com genes facilitate DNA binding and transportation into the cytoplasm. In N. gonorrhoeae, ComE binds the transforming DNA in the periplasm and is necessary for efficient uptake, ComL is directly or indirectly involved in puncturing the peptidoglycan layer, and ComA is an inner membrane protein that helps transport DNA into the cytoplasm (21). Once inside the cytoplasm, X. fastidiosa has the machinery necessary for RecA-mediated recombination.

Table 2.

Annotated com genes present in the X. fastidiosa strain Temecula genome

Locus	Gene name	Putative function of gene product
0031	comJ	Transformation competence-related protein
0042	comF	Competence protein F
0358	comA	DNA uptake protein
0464	comM	Competence-related protein
1558	comE	DNA transport competence protein
1756	comL	Competence lipoprotein

Initially, it was thought that most plant-pathogenic bacteria, such as Pseudomonas syringae, were primarily clonal (38). However, recent studies have provided evidence indicating that limited and biased sampling within populations may have underestimated the rate of homologous recombination in these organisms (20, 44). In addition, several highly specialized pathogens recombine at much lower frequencies than those of their nonspecialized relatives, and it has been hypothesized that the emergence of new pathogens could be characterized by high recombination rates during adaptation to new environments or hosts, followed by a drop in recombination frequency once the pathogen has settled into its new niche (15). Recent studies have shown that X. fastidiosa's genome is subject to recombination and that this process could play a large role in generating genetic diversity and affect the organism's evolution (1, 39, 45). It has also been hypothesized that Pierce's disease-causing strains of X. fastidiosa in the United States, which have little allelic diversity, diverged from an isolate introduced from Costa Rica (34). If true, these recently diverged strains could rapidly develop novel genotypes as they recombine with strains from areas of endemicity, though this has not yet been shown to happen. Here, evidence has been provided showing that X. fastidiosa is naturally competent and that transformation is a likely mechanism for DNA to be horizontally transferred between organisms. Further studies of X. fastidiosa's natural competency could provide insight into the effects of recombination on pathogen diversity and the emergence of new diseases.