ABSTRACT
Phage tail-like bacteriocins (tailocins) are bacterially produced protein toxins that mediate competitive interactions between cocolonizing bacteria. Both theoretical and experimental research has shown there are intransitive interactions between bacteriocin-producing, bacteriocin-sensitive, and bacteriocin-resistant populations, whereby producers outcompete sensitive cells, sensitive cells outcompete resistant cells, and resistant cells outcompete producers. These so-called rock-paper-scissors dynamics explain how all three populations occupy the same environment, without one driving the others extinct. Using Pseudomonas syringae as a model, we demonstrate that otherwise sensitive cells survive bacteriocin exposure through a physiological mechanism. This mechanism allows cells to survive bacteriocin killing without acquiring resistance. We show that a significant fraction of the target cells that survive a lethal dose of tailocin did not exhibit any detectable increase in survival during a subsequent exposure. Tailocin persister cells were more prevalent in stationary- rather than log-phase cultures. Of the fraction of cells that gained detectable resistance, there was a range from complete (insensitive) to incomplete (partially sensitive) resistance. By using genomic sequencing and genetic engineering, we showed that a mutation in a hypothetical gene containing 8 to 10 transmembrane domains causes tailocin high persistence and that genes of various glycosyltransferases cause incomplete and complete tailocin resistance. Importantly, of the several classes of mutations, only those causing complete tailocin resistance compromised host fitness. This result indicates that bacteria likely utilize persistence to survive bacteriocin-mediated killing without suffering the costs associated with resistance. This research provides important insight into how bacteria can escape the trap of fitness trade-offs associated with gaining de novo tailocin resistance.
IMPORTANCE Bacteriocins are bacterially produced protein toxins that are proposed as antibiotic alternatives. However, a deeper understanding of the responses of target bacteria to bacteriocin exposure is lacking. Here, we show that target cells of Pseudomonas syringae survive lethal bacteriocin exposure through both physiological persistence and genetic resistance mechanisms. Cells that are not growing rapidly rely primarily on persistence, whereas those growing rapidly are more likely to survive via resistance. We identified various mutations in lipopolysaccharide biogenesis-related regions involved in tailocin persistence and resistance. By assessing host fitness of various classes of mutants, we showed that persistence and subtle resistance are mechanisms P. syringae uses to survive competition and preserve host fitness. These results have important implications for developing bacteriocins as alternative therapeutic agents.
INTRODUCTION
Diverse microbes inhabit shared environments and compete for limited resources. Competition for these resources can be mediated by the secretion of toxins, such as antibiotics, type VI effectors, and bacteriocins, that enable the producing cells to maintain their dominance (1, 2). Bacteriocins are bacterially produced protein toxins that are lethal toward strains related to the producer (3, 4). These antibacterial toxins have been proposed for several decades as antibiotic alternatives to treat or prevent infection spread in both humans and plants (5, 6), in addition to being used in food preservation (7). Given their ubiquitous nature, where sequenced bacterial genomes are commonly predicted to encode at least one bacteriocin (8–10) and bacteria isolated from diverse environments often produce detectable bacteriocins (e.g. references 8, 11, 12), it is reasonable to predict that bacteriocin-producing populations exert a selective force on cocolonizing sensitive populations to either gain resistance or avoid killing through a different mechanism. Resistance evolution, often involving a heritable mutation in either the toxin receptor or membrane translocator genes, is a common mechanism to avoid being killed (13, 14). Resistant mutants, however, are likely to suffer fitness costs associated with the mutation, which reduces their ability to proliferate in the environment. Therefore, resistant mutants are competitive only in environments where there is substantial or sustained bacteriocin exposure; otherwise, they are competitively inferior to their sensitive progenitors. Conversely, bacteriocin-producing populations are dominant over sensitive populations but are competitively inferior to resistant populations, as the result of the resources wasted on ineffective toxins. These nontransitive dynamics underlie a rock-paper-scissors model for microbial competition (13–15), where a community composed of all three genotypes is maintained as a result of negative frequency-dependent selection (14). However, these dynamics, which are built primarily on modeling and laboratory culture-based experiments, are dependent on a small number of qualitative states. It is unknown whether or how quantitative resistance or physiological tolerance in otherwise sensitive populations will affect community competitive dynamics. The ability of bacteriocin-sensitive cells to persist under toxin pressure has important implications for the ecology of microbes generally and for the potential to use bacteriocins as therapeutics specifically. For instance, previous studies have reported that, although sensitive cells may not coexist with the producing strain in a well-mixed environment (16–18), they still prevail in the competitive in vivo environments (13, 14, 19). Little is known, however, regarding the mechanism(s) that allows the maintenance of a sensitive population despite sustained bacteriocin exposure.
Bacteriocins are classified into different groups based on their structure, composition, and mode of action. Tailocins are bacteriocins that resemble bacteriophage tails and are grouped into R-type (with a retractile core tube) and F-type (flexible) (20, 21). In the opportunistic human pathogen Pseudomonas aeruginosa, other environmental pseudomonads, and Burkholderia sp., phage-derived tailocin bacteriocins were shown to antagonize competitors, including pathogenic strains (22–28). In fact, a recent study by Principe et al. suggested the effectiveness of foliar sprays of tailocins produced by Pseudomonas fluorescens in reducing the severity and incidence of bacterial spot disease in tomato caused by Xanthomonas vesicatoria (28). Other studies have also indicated the potential use of engineered R-type tailocins in suppressing foodborne pathogens using in vivo models (29, 30). Our group has previously characterized an R-type tailocin from a plant-pathogenic bacterium, Pseudomonas syringae pv. syringae B728a (31). This R-type tailocin showed antagonistic potential against several pathovars of P. syringae that cause serious diseases and substantial losses in economically important crops, such as common bean (P. syringae pv. phaseolicola [Pph]), soybean (pv. glycinea), chestnut (pv. aesculi), and kiwifruit (pv. actinidiae) (31). A broad spectrum of tailocin-mediated antagonistic interactions in P. syringae have been described recently (11). Yet, we have a very limited understanding regarding the defense responses against tailocin by the target pathogen, which is key information for designing tailocins as therapeutic agents.
Tailocins are considered to be potent killers, as a single tailocin particle is predicted to kill a sensitive cell and an induced producer cell can release as many as 200 particles (32, 33). R-type tailocins, once bound to the cell surface receptors of the target cells, puncture through the cell membrane and cause cell death by membrane dissipation (3, 32). Specific lipopolysaccharide (LPS) components of the target cells are known to serve as surface receptors of tailocins (34–36). LPS is composed of three components, namely, the lipid A, core oligosaccharide, and the O-polysaccharide (O-antigen) (37, 38). Although the lipid A and core region are mostly conserved within a species, the O-antigen region varies extensively in its chain length and composition of sugars (39). Biosynthesis and transport of LPS to the outer membrane as well the modification of O-antigen require complex processes involving a number of highly diverse genes (38). Little is known about the extent of LPS modification required for tailocin resistance and persistence and the consequence of these modifications in host fitness and virulence in the target pathogen.
This study aimed to examine the phenomena that addresses both of the following crucial questions related to bacteriocins: (i) how do sensitive cells survive lethal bacteriocin exposure and (ii) can bacteriocins be used as stand-alone control measures? We demonstrate that a sensitive population can employ multiple strategies to survive toxin exposure without acquiring an otherwise fitness-reducing mutation. One strategy was physiological persistence, a mechanism that enables a subpopulation of sensitive cells to transiently survive lethal doses of the bacteriocin without acquiring genetic changes. The second strategy relies on acquiring subtle genetic changes (incomplete resistance) that do not impose a detectable fitness cost, while still allowing the mutants to better survive bacteriocin exposure. Finally, we demonstrated that the complete resistance mutants suffered a fitness cost within a susceptible host plant, whereas both high persister mutants and incomplete resistant mutants were equally fit compared with the wild type. This work suggests that bacterial cells can employ mechanisms to survive antagonistic toxins but still preserve their host colonization potential. This work has important implications for how bacteria can potentially avoid rock-paper-scissors dynamics that is widely understood to be important in mediating interactions between toxin-producing, sensitive, and resistant genotypes and in using bacteriocins as potential therapeutic agents.
RESULTS
A subpopulation of P. syringae pv. phaseolicola (Pph) cells survives tailocin exposure by persistence that increased in the stationary state.The activity of the purified tailocin was determined to be 103 to 104 activity units (AU) and 1.25 × 107 to 4.25 × 109 lethal killing units ml−1. The MIC was determined to be 100 AU when exposed to ∼106 viable target cells at their logarithmic growth. No loss of tailocin activity was observed for a period of over 6 months in the buffer (10 mM Tris [pH 7.0] and 10 mM MgSO4) at 4°C.
Purified tailocin was used to test its killing effects on stationary and log-phase cultures of the Pph target cells in a broth environment. After an hour of a 100-AU tailocin treatment, a consistent reduction (3.59 ± 0.12 log) in the viable population occurred for logarithmic cultures, while a significantly lower reduction (1.38 ± 0.14 log) occurred for the stationary cultures. Further analysis showed that, upon treatment of an equivalent number of viable cells, stationary cells consistently survived 10- to 100-fold more than the logarithmic cells (Fig. 1A; see Fig. S1 in the supplemental material). Surviving colonies, especially those from the stationary phase, were predominantly sensitive upon tailocin reexposure, suggesting survival by a persistence mechanism (see below).
Differences in tailocin survival and proportion of surviving phenotypes between stationary- and log-phase cultures of Pph. (A) Cultures were treated with a lethal dose of tailocin (100 AU), and viable cells before and after 1 h of treatment were enumerated. (B) Tailocin death curve of the two cultures of Pph upon exposure to tailocin (100 AU). Viable populations were enumerated before and after 1, 4, 8, and 24 h of treatment (see Fig. S2). Three independent experiments were performed with 3 to 6 biological replicates per time point. The mean and standard error of the mean are graphed. A P value of < 0.05 or different letters indicate significant differences between stationary- and log-phase treatments for a given time point, as analyzed in SAS 9.4 with proc Glimmix. (C) Percentage of surviving colony phenotypes upon tailocin retreatment. Randomly selected surviving colonies (n = 12 to 44 for each growth phase and hours of treatment from three independent experiments) were subcultured and treated again with tailocin, and the percentage of the surviving phenotype was calculated.
The persistent subpopulation is maintained under prolonged exposure time and increased concentration of tailocin.Tailocin treatments were applied to both stationary and log cultures of Pph for up to 24 h with enumeration of surviving population before and after 1, 4, 8, and 24 h of tailocin treatment to generate a tailocin death curve. After a steep reduction in the population within the first hour of treatment, further killing of the cells that survived the first hour of treatment did not occur in either culture (Fig. 1B). Twenty-four hours posttreatment, although the overall population increased (Fig. 1B), individual treatments showed different results; for some replicate treatments, the population remained constant, suggesting maintenance of the persistent state, while for some other replicates, the population increased due to a division of cells that acquired tailocin resistance (see Fig. S2).
Upon tailocin retreatment, >90% of stationary and >60% of log cells that survived the first hour treatment were as sensitive as the wild type (i.e., persistent) as in Fig. 1C. The proportion of persistent survivors was higher in the stationary cultures than in the log cultures at all time points (Fig. 1C). Tailocin persistent cells were recovered from both cultures even after 24 h of tailocin treatment, although the proportion decreased over time (Fig. 1C). Tailocin activity was detected in the supernatants recovered from the treated samples that contained persistent cells (see Fig. S3), confirming saturation of tailocin in the treatment. Although a slight reduction of activity was observed when the tailocin preparation was mixed with undiluted stationary supernatant compared with the log supernatant, no difference was detected upon diluting the supernatants by 1,000- to 20,000-fold before mixing with tailocin (similar to how cultures were diluted for tailocin treatment) (see Fig. S4). This suggested that the increased tailocin persistence in the stationary phase is not related to inhibition of tailocin activity by an extracellular component.
Upon treating the cells with a concentrated tailocin (900 AU), the surviving population decreased such that no difference in survival between the stationary and log cultures was detected (Fig. 2A). However, even with this higher level of tailocin applied, the proportion of tailocin persistent cells remained higher for stationary-phase survivors than that for the log-phase survivors (Fig. 2B).
Dynamics of tailocin survival with concentrated tailocin treatment. (A) Cultures were treated with a high dose of tailocin (900 AU), and viable populations pre- and posttreatment were determined. Experiments were repeated at least three times with 3 to 6 biological replicates per time point. The mean and standard error of the mean are graphed. A P value of <0.05 indicates significant difference within grouped bars, as analyzed in SAS 9.4 with proc Glimmix. (B) Percentage of surviving colony phenotypes after treatment with concentrated (900 AU) tailocin for 1 hour. Although most of the surviving colonies were either incomplete resistant or resistant, persistent cells were maintained even at high tailocin concentration. Surviving colonies were tested during three independently repeated experiments for both cultures.
Tailocin exposure selects for heritable mutants showing increased persistence and heterogenous resistance.In addition to the recovery of the tailocin persistent subpopulation, we recovered an unique mutant, referred to here as high persistent-like (HPL), which showed a significantly higher survival level than the wild type of broth treatment (Fig. 3A). However, long-term exposure of this strain within overlay conditions showed similar sensitivity to wild type (Fig. 3B). Furthermore, the HPL phenotype did not differ in survival between the stationary and log phases even at a higher concentration of tailocin (Fig. 4). Next, the category of mutants recovered were conditionally sensitive and are referred to here as incomplete resistant (IR) mutants (see Fig. 1C and 2B for proportions). These mutants lost sensitivity in the broth even at high tailocin concentrations (Fig. 3A) but displayed some sensitivity in the overlay (Fig. 3B). Lastly, complete tailocin resistant (R) mutants that were insensitive to tailocin under both treatment conditions were also recovered. Of the four complete resistant mutants we selected, two (R1 and R4) showed an unique rough colony morphotype.
Treatment response of tailocin persistent and resistant lines. (A) Reduction in the population of tailocin persistent and resistant mutant lines upon retreatment with tailocin. Log cultures of each line were treated with 900 AU of tailocin, and the change in the population was calculated after an hour of tailocin exposure. At least three separate colonies of each line were tested, and experiments were repeated a minimum of three times. Means of the difference in log-transformed viable population pre- and posttreatment are graphed. Error bars indicate the standard error of the mean. (B) Assessment of the response of mutant lines to tailocin in overlay conditions. Dilutions of tailocins (shown on the leftmost column) were spotted over the culture lawn of each of the lines. The yellow line indicates the dilution up to which visible killing was consistently observed. HPL, high persistent-like; IR, incomplete resistant; R, resistant.
Dynamics of tailocin survival for the high persistent-like (HPL) mutant. Cultures were treated with various tailocin doses, and viable cells pre- and posttreatment were enumerated. Means and standard errors from three independently repeated experiments with at least three biological replicates per experiment are reported. More killing occurred at high tailocin concentrations, and no difference in stationary and log phases was observed. P values of >0.05 indicate no significant differences within grouped bars, as analyzed in SAS 9.4 with proc Glimmix.
Mutations involve various genes likely associated with LPS biogenesis and modification.Genome sequencing and variant identification of the high persistent-like (n = 1), incomplete resistant (n = 6), and resistant mutants (n = 4) were performed by mapping the Illumina reads with the parental reference sequence. Mutants isolated in different experiments showed mutations in different loci. A specific LPS biogenesis region (Fig. 5A) was identified in the Pph genome that showed the most prominent role in tailocin resistance/sensitivity. The HPL mutant contained a 16-bp deletion that caused a frameshift near the 3′ end of a hypothetical protein-coding gene cotranscribed with the LPS genes. No functional evidence could be found for the hypothetical protein by in silico analysis except that it is predicted to contain a signal peptide domain at its N-terminal and 8 to 10 transmembrane domains. Most genes identified for complete and incomplete resistance encode glycosyltransferases and related proteins that are likely involved in LPS biogenesis (Table 1; Fig. 5A). For example, for the gene PSPPH_0957 that encodes a glycosyltransferase, a homolog of WbpX, a missense mutation that caused a frameshift and premature termination (R1 mutant), and a missense mutation (R2 mutant) in the middle of the gene caused complete resistance, while insertion of a few bases at the 3′ end of this gene caused incomplete resistance (IR4 mutant). In the same region, a missense mutation in the middle of a gene (PSPPH_0960) encoding a glycosyltransferase family protein 2 caused incomplete resistance (IR3 mutant). Moreover, one of the incomplete resistant mutants (IR6) showed mobilization of the 100-bp miniature inverted-repeat transposable element (MITE) sequence present in the Pph genome, as described by Bardaji et al. (40). The MITE insertion likely inactivated PSPPH_0963, a gene that is annotated to encode a flavin adenine dinucleotide (FAD)-dependent oxidoreductase. Among other mutants, R3 contained a missense mutation in PSPPH_0952, the closest homolog of which in P. aeruginosa is predicted to encode a Rmd protein involved in GDP-d-rhamnose synthesis. The R4 mutant had a frameshift in PSPPH_0983 that encodes a homolog of WbpY, likely involved in the extension of α 1-2 glycosyl linkages of the O-antigen. Other genes identified for IR1 (PSPPH_3226), IR2 (PSPPH_2810), and IR5 (PSPPH_0520) were glycosyltransferases likely involved in the extension of main or side chains of the O-antigen. Some of these LPS regions (e.g., PSPPH_0946-0966, PSPPH_0983, PSPPH_2810, and PSPPH_3226) appeared to be associated with remnant mobile elements, which may suggest their acquisition via horizontal transfer or their modification via internal mobilization.
Genomic mapping of mutations and in planta fitness of selected mutants. (A) Map of predicted lipopolysaccharide O-antigen biogenesis region (PSPPH_0945 to PSPPH_0966) of Pph that showed pronounced effects in tailocin persistence and resistance. Refer to Table 1 for a complete list of genes at this and other genomic locations. (B) Strains and mutants were syringe infiltrated into green bean primary leaves. Experiments were repeated at least twice with 8 biological replications per strain. A representative experiment is presented. Error bars indicate standard error of the mean. Different letters indicate a significant difference (P < 0.05) for a given time point. Pph, wild-type strain; ΔhrpL::Pph, Pph ΔhrpL type III secretion system mutant; HPL, high persister-like; IR, incomplete resistant; R, complete resistant.
Tailocin sensitivity, predicted loci of mutation, and effect on target gene of tailocin high persistent and resistant mutants and selected allele-swapped strains
The mutant phenotypes detected by genome sequencing were further confirmed by Sanger sequencing of the target region as well as by swapping the mutant alleles to the wild-type background and vice versa of the selected mutants (HPL, IR4, R1, R3, and R4). In all cases, the allele-swapped strains showed the phenotypes expected (Table 1). This confirmed that the mutations identified by genome sequencing were responsible for the tailocin high persistent-like and resistant phenotypes.
LPS analysis of the mutants and the wild-type Pph showed that the complete resistant mutants lacked a fully formed O-antigen region, whereas the high persistent-like and majority of incomplete resistant mutants still possessed the O-antigen, with minor to undetectable changes. One of the incomplete resistant mutants (IR4), however, showed a very different and faint O-antigen band (see Fig. S5).
In planta fitness is compromised for complete resistant mutants which are devoid of the LPS O-antigen.High persistent-like and incomplete resistant mutants did not suffer any fitness cost when infiltrated into a compatible host (green bean), whereas the complete resistant mutants did (Fig. 5B; see Fig. S6). Up to a 100-fold reduction in the population was seen for the complete resistant mutants (equivalent to the reduction in the type III secretion system mutant ΔhrpL::Pph). At 48 hours postinfection (hpi), although the complete resistant mutant population increased compared with the type III mutant, it was still significantly lower than wild type, HPL, and IR mutants (Fig. 5B; Fig. S6). These results suggest that persistence and incomplete resistance are mechanisms used by P. syringae to survive attack by competitor strains while maintaining host fitness and virulence potential.
DISCUSSION
There have been renewed research interests in alternative treatment strategies for bacterial pathogens due mainly to the growing threats of antibiotic-resistant infections (5, 7). Bacteriocins, including tailocins, have long been proposed as effective and more specific alternatives to broad-spectrum antibiotics (5, 6). However, a critical question that remains unaddressed for using bacteriocins as pathogen control agents is how does the sensitive pathogen population respond to the application of lethal doses of bacteriocins? Moreover, although it is predicted in the rock-paper-scissors model of bacteriocin-mediated interaction that sensitive cells coexist with the producer through ecological mechanisms (13, 14, 19), there is no description of the possibility of maintaining a bacteriocin-sensitive population through a physiological mechanism.
In this study, we addressed these questions using a phage tail-like bacteriocin (i.e., tailocin) produced by P. syringae pv. syringae strain B728a in killing target cells of P. syringae pv. phaseolicola 1448A. We showed that, upon exposure of a lethal dose of tailocin, a subpopulation of sensitive cells survives without undergoing genetic changes. The fraction of this subpopulation, termed here as tailocin persistent subpopulation, increased significantly in the stationary phase compared with the logarithmic phase of growth. Similar results of increased persistence in the stationary phase and biofilms with antibiotics have been previously described in other bacteria, including P. aeruginosa (41, 42). By repeated exposures of this subpopulation to the same or higher doses of tailocin, we showed that they have not gained any heritable resistance and that physiological persistence is the only mechanism for their survival. Moreover, a prolonged tailocin exposure generated a killing pattern similar to that reported for persistent subpopulation upon antibiotic treatment (43, 44). Persistence was maintained for at least 24 h with tailocin exposure, a phenomenon that was more evident in some treatment replicates in which resistant evolution did not occur (see Fig. S2 in the supplemental material). Although increasing the tailocin concentration killed some of the persistent survivors and the difference in survival between the two growth phases was no longer seen, stationary phase-derived cells still exhibited a higher percentage of persistence survivors than the log-phase cells upon reexposure. This indicated that the probability of a successful hit in stationary phase is lower than in log phase. Since tailocins are thought to be target cell-specific and are not known to have off-target effects, a higher concentration of tailocin could be used to achieve a more effective pathogen control. However, although at a low level, persistence was still maintained even with high-dose tailocin treatment, and an inherent emergence of either complete or incomplete resistance was frequently observed. As such, although a significant reduction in the pathogen population and disease pressure can be obtained with tailocins, a stand-alone tailocin treatment might not be enough to achieve a sustainable pathogen control.
The use of the term persistence in relation to antimicrobial survival is disputed to some extent and is sometimes used interchangeably with “tolerance” and “viable but not culturable state.” Here, we used the term persistence, as this phenotype was seen only in a subpopulation and resulted in a biphasic death curve and cells resuscitated almost immediately upon tailocin removal and were equally sensitive to the wild-type cells upon reexposure. This definition of persistence has been suggested previously (44). Persistence to antimicrobials is being increasingly recognized for its role in antimicrobial treatment failures with bacterial infections (45). Various mechanisms are implicated in the maintenance of persistence (46, 47), although persistence responses can be different based on the stresses involved and their mode of action (48). Of these mechanisms, toxin-antitoxin (TA) systems are the most studied in terms of formation of persister subpopulations (49–51). TA systems were shown to be induced when cells were starved for certain sugars and amino acids or by exposure to osmotic stresses that altered ATP levels in the cell (48, 52). However, it was also shown that activation of a TA system does not always induce persister formation (52). Additionally, recent findings have indicated a mechanism mediated by the guanosine penta- or tetraphosphate (ppGpp) for persister formation that is not dependent on a TA system (53). A strong stationary-state effect, which likely involved a starvation response, was shown to increase persistence by 100- to 1,000-fold in Staphylococcus aureus with ciprofloxacin treatment (52). Whether similar mechanisms of TA and independent ppGpp systems regulate tailocin persistence or a specific mechanism for tailocin and/or related bacteriophage exist remain to be determined. Nevertheless, our data of the difference in tailocin persistence between the stationary-phase and log-phase cultures suggest that metabolic inactivity and starvation-induced stress could be strong factors in tailocin persistence.
Few previous studies have demonstrated growth phase-dependent differences in LPS O-antigen chain length and composition or their regulatory pathways (54, 55). In a previous study with P. fluorescens, exponentially growing cells had a significantly higher rate of cell lysis than stationary- or decline-phase cells with bacteriophage PhiS1 (56). Moreover, a recent study showed that sensitivity to colicin, a bacteriocin from Escherichia coli, could be altered by growth condition-dependent LPS O-antigen changes (57). Growth phase-dependent LPS modification, although it has not been reported so far in P. syringae, could be contributing to the difference in tailocin persistence. We also observed that undiluted stationary-phase supernatant inhibited tailocin activity to some extent compared with the log-phase supernatant. However, whether this is linked to differences in the secreted LPS between the two supernatants or other cellular factors needs further assessment. Moreover, we also demonstrated that mutation in one of the hypothetical proteins containing a signal peptide and several transmembrane domains caused increased tailocin persistence. Since the gene for this hypothetical protein occurs in the same operon as other LPS biogenesis genes, it is likely that it plays a role in O-antigen biogenesis and/or modification, potentially reducing tailocin interaction with the cells. A protein BLAST search showed that only a subset of P. syringae strains (including pathovars actinidiae, aesculi, and morsprunorum) and some environmental bacteria, such as the genus Burkholderia, carry the homologs in their genome. Further experiments focusing on the mechanistic details of the protein will be required for determining its role in tailocin persistence.
Phase variation is another mechanism that is known to cause increased survival to surface active antimicrobials (e.g., phages and host immune defenses) (58). Phase variation is a gene regulation system that induces heterogenous expression of specific genes in a clonal population (58–60). Although phase variation is heritable, the “on” and “off” switch from the variant to wild-type phenotype occurs randomly, amounting to 10−4 to 10−1 per generation, which is significantly greater than what is expected by mutational events (61). Phase variation has been shown to modify LPS operons in Salmonella enterica spp. (62), P. aeruginosa (63), and P. fluorescens (64). In S. enterica serovar Typhimurium, phase-variable glycosylation of the O-antigen was reported to cause temporal development of phage resistance (65). Other reports of the role of phase variation in defense against bacteriophages, which also target LPS, are available (66, 67), including in the closely related P. aeruginosa (63). Therefore, phase variation in tailocin survival cannot be ruled out since LPS is the receptor involved. Since the rate of switch from the tailocin surviving variant to the sensitive wild type upon tailocin removal will have to be much greater than that reported (61), the tailocin persistent phenotype is unlikely to be the result of phase variation. It remains to be determined, however, if the incomplete resistance could be caused by phase-variable gene switching. In one of the incomplete resistant mutants, the target gene was modified by movement and integration of an internal mobile genetic element (MGE). MGEs have been shown to cause phase variation in other bacterial systems (68, 69). A high-throughput sensitivity screening of the mutant progeny colonies will be required to confirm if the tailocin resistant phenomena involve a phase-variable gene switch.
To our knowledge, the persistent as well as incomplete resistant phenotypes observed here have not been described before with bacteriocins. Here, using a Pseudomonas syringae tailocin, we showed that, in addition to the persistent subpopulation, resistant lines with various degrees of sensitivity are selected by exposure of target cells to tailocins. Previous studies in P. aeruginosa and other bacteria also showed that LPS serves as the receptor to tailocin and other bacteriocins (34, 36, 57). LPS analysis and in planta fitness assessment showed that to gain complete tailocin resistance, cells have to lose their LPS O-antigen. This comes with a fitness trade-off. On the contrary, by maintaining a persistent subpopulation and/or by undergoing subtle genetic changes, the LPS integrity was preserved for the most part and fitness within the host was maintained. Similar results of reduced in planta virulence and fitness were shown in phage resistant mutants devoid of O-antigen (70). Moreover, two of the complete resistant mutants (R1 and R4) in our study showed a typical rough-colony morphotype. Rough-colony morphotypes lacking the LPS O-antigen were resistant to phages and had reduced in planta virulence, as described previously in P. syringae pv. morsprunorum isolates from plum and cherry (64). In other plant-pathogenic bacteria, such as Xanthomonas oryzae pv. oryzae and Xylella fastidiosa, loss of the O-antigen reduced type III secretion into plants (71) and increased recognition of the pathogen by the host immune response (72), respectively. In both cases, plant virulence of the mutants was significantly reduced (71, 72). Also, although the incomplete resistant mutants contained mutations in the LPS biogenesis genes and mostly had their O-antigen region intact, they did not show any fitness trade-offs in our growth chamber experiments. However, under a natural environment that involves more severe environmental stresses, it could be expected that their fitness would be altered. Under these circumstances, the persister subpopulation, which does not involve any genetic changes, would enable the target pathogen to withstand the host defense while surviving the competition and ensure that its lineage is maintained.
Taken together, both previous reports and our results suggest that full resistance to tailocin incurs a significant fitness cost. Our work demonstrates that persister and incomplete resistant subpopulations of the sensitive strain preserve their host fitness despite suffering bacteriocin exposure. These results have important implications for the mechanisms and ecological processes that can promote the coexistence of both sensitive and bacteriocin-producing populations. In Fig. 6, we present a model of the three cell types (resistant, incomplete resistant, and persister) that can be detected following bacteriocin-mediated selection. In particular, we view the persister subpopulation as one that can survive bacteriocin exposure, without paying any long-term fitness cost. In the short term, however, we believe that persister cells have significantly reduced fitness because they are likely not dividing (see Fig. S2). Thus, bacteriocin persisters may best be thought of as a means to switch, at high frequency, between being phenotypically resistant and phenotypically sensitive.
A visual representation of the various tailocin surviving subpopulations and their phenotypes upon exposure to tailocin.
MATERIALS AND METHODS
Bacterial strains, media, and culture conditions.All bacterial strains, plasmids, and mutants used in this study are listed in Table S1 in the supplemental material. P. syringae pv. syringae wild-type (WT) strain B728a (73) and its tailocin-defective mutant ΔRrbp (31) were used to prepare the treatment supernatants. P. syringae pv. phaseolicola (Pph) 1448A (74) was used as the target strain. Tailocin high persistent-like, resistant, and selected complemented strains of Pph generated in this study are described in Table 1. King’s medium B (KB) was used to culture the strains. Liquid cultures were prepared by inoculating individual colonies from a 2-day-old KB agar plate into 2 ml of liquid medium at 28°C with shaking at 200 rpm. Antibiotics kanamycin (Km), chloramphenicol (Cm), tetracycline (Tet), gentamycin (Gm), rifampin (Rif), nitrofurantoin (NFT), and nalidixic acid (NA) were used at 50, 25, 10, 50, 50, 50, 50 μg/ml final concentrations, respectively.
Tailocin induction and purification.Purified tailocin and control treatments were prepared from logarithmic (log) cultures of P. syringae pv. syringae B728a and ΔRrbp, respectively, using a polyethylene glycol (PEG) precipitation protocol as previously described (31, 75). Briefly, 100-fold diluted overnight B728a cultures were subcultured for 4 to 5 h in KB broth before inducing with a 0.5-μg ml−1 final concentration of mitomycin C (GoldBio). Induced cultures were incubated for 24 h with shaking at 28°C. Next, cells were pelleted by centrifugation, and the supernatants were mixed with 10% (wt/vol) PEG 8000 (FisherScientific) and 1 M NaCl. Supernatants were then incubated either in ice for 1 h and centrifuged at 16,000 × g for 30 min at 4°C or incubated overnight at 4°C and centrifuged at 7,000 × g for 1 hour at 4°C. Pellets were resuspended (1/10 to 1/20 of the original volume of the supernatant) in a buffer (10 mM Tris [pH 7.0] and 10 mM MgSO4). Two extractions with equal volumes of chloroform were performed to remove residual PEG. Tailocin activity was confirmed by spotting dilutions of 3 to 5 μl of both the tailocin and control supernatants onto soft agar overlays of Pph. The relative activity of tailocin was expressed as arbitrary units (AU), as obtained from the reciprocal of the highest dilution that exhibited visible tailocin killing in an overlay seeded with ∼108 CFU of Pph log-phase cells. Lethal killing units of the purified tailocin were determined using a Poisson distribution of the number of surviving colonies after treatment with different dilutions of the tailocin, as described previously (27, 76).
Tailocin treatment and survival assessment for stationary-phase and log-phase cultures.To assess tailocin activity against the stationary and log phases of Pph, individual colonies growing on KB agar plates for ∼2 days were inoculated into 2 ml of liquid KB medium. Following incubation at 28°C with shaking at 200 rpm overnight, the cultures were back diluted 1,000-fold into fresh KB. The back-diluted cultures were either incubated for 28 to 30 h to prepare stationary cultures or back diluted 100-fold at 24 h and cultured for another 4 to 6 h to prepare log cultures (see Fig. S7 in the supplemental material for a growth curve of Pph).
Stationary cultures were diluted 20,000-fold, and logarithmic cultures were diluted 1,000-fold (∼105 to 106 CFU/ml for both cultures) (Fig. 1A) in fresh KB before tailocin treatment. Treatment was applied by mixing 10 μl of diluted cultures in 90 μl of purified tailocin diluted in KB. After treatment, samples were incubated for ∼1 h at 28°C and washed twice to remove residual tailocin particles. Washing was performed by mixing the treated culture in 900 μl of fresh KB, followed by centrifugation at 12,000 × g for 2 min. The top 900-μl fraction was discarded, and the bottom 100-μl fraction was serially diluted and either spread or spot plated to enumerate the surviving population. Plates were incubated at 28°C for 1 to 3 days before enumeration. Serial dilutions of both stationary-phase and log-phase cultures were spotted onto KB agar to enumerate the untreated population. Experiments were performed with various tailocin concentrations (i.e., 100 AU, 500 AU, and 900 AU).
Tailocin retreatment to differentiate persistence and resistance.Surviving colonies were treated again with tailocin to differentiate them into persistent or resistant colonies. Retreatment was performed by an overlay method as described previously (75) or by broth treatment as discussed above. The overlay method was used to determine the AU of the tailocin preparation with the selected mutant lines. Broth exposure was used to the calculate reduction in the population of log-phase cultures after treatment. Surviving colonies were differentiated into various phenotypes as follows: persistent (sensitive to tailocin to the wild-type level in both the overlay and broth method), high persistent like (completely sensitive in the overlay but survived significantly more than the wild type under broth conditions), incomplete resistant (showed conditional sensitivity [i.e., some sensitivity in overlay but no significant sensitivity in the broth]), and complete resistant (were insensitive under both conditions).
Time-dependent death curve with tailocin treatment.Prolonged tailocin exposure was performed with both the stationary- and log-phase cultures of Pph to generate death curves. Treatment was applied in a 96-well plate as before. Surviving populations were enumerated at 1, 4, 8, and 24 h following treatment, and randomly selected surviving colonies were reexposed to tailocin to differentiate them into persistent or resistant phenotypes.
Tailocin recovery from the treated samples and activity testing.Stationary-phase and log-phase cultures treated with purified tailocin for 1 to 24 h as described above were centrifuged, and the supernatant was collected and filter sterilized using a 0.22-μm syringe filter. Supernatants were diluted 5-, 10-, 50-, and 100-fold in KB and spotted onto Pph overlay. Purified tailocin particles diluted in KB were also included as a control treatment.
Determining the effect of stationary- and log-phase supernatant on tailocin activity.Stationary- and log-phase cultures were prepared as described above by culturing Pph cells in KB broth for either 28 to 30 h or for 4 to 6 h, respectively. Cultures were centrifuged for 2 min at 12,000 × g, and the supernatant was filter sterilized using a 0.2-μm syringe filter. Stationary- and log-phase supernatants were diluted 1,000- to 20,000-fold in KB (according to how the cultures were diluted for tailocin treatment). Various dilutions (10-, 50-, 100-, and 1,000-fold) of purified tailocin were prepared in the stationary- and log-phase supernatants. Dilutions were spotted on a Pph lawn using the overlay method.
Genome sequencing and analysis of the tailocin high persistent-like and resistant mutants.The tailocin high persistent-like (HPL) and complete and incomplete resistant mutants recovered from tailocin treatment of Pph wild-type cells (and confirmed by retreatment) were selected for genome sequencing. DNA was extracted from the overnight cultures with the Promega Wizard genomic DNA purification kit using the manufacturer’s protocol. DNA quantity and quality were assessed with a Qubit 3 fluorometer using the Qubit double-stranded DNA (dsDNA) high-sensitivity (HS) assay kit (Thermo Fisher Scientific) and Nanodrop 2000 (Thermo Scientific). A uniquely indexed library from each mutant line was prepared using the DNA flex kit (Illumina). An approximately equimolar pool of libraries was generated, and 150-bp paired-end reads were sequenced on an Illumina MiSeq at the Penn State Genomics Core Facility.
Resulting forward and reverse read files were paired and mapped to the Pph chromosomal and plasmid sequences using Geneious R10.2 with default parameters for medium sensitivity. Next, genetic variants were identified using “Find variations/SNPs” program within Geneious using default settings. Regions supported by a minimum of 10 reads and >90% variant frequency were selected. Moreover, variants shared among all the mutants that were generated in independent experiments were discarded as misalignments. Variants identified by Geneious were confirmed in the contigs generated by de novo assembly of the paired reads with SPAdes 3.11.0 using k-mer sizes of 21, 33, 55, 77, 99, and 127 with careful mode selected to minimize mismatches and short indels. Indels were also detected and visualized in the contigs by Harvest suite tools (77). Sequencing generated 1,986,400 to 2,857,324 paired reads per genome with 48× to 69× genome coverage. De novo assembly generated 308 to 338 contigs per genome, with N50 values of 75,222 to 81,220 bp. The total assembly size was 5.97 Mbp with a G+C content of 57.96%, which are values similar to those of the Pph reference genome (74).
Presence-absence and similarity searches of the selected genomic regions and genes implicated in tailocin persistence and resistance were performed with NCBI and IMG-JGI databases with the BLAST algorithm using the Pph sequences as the query. InterProScan (78) and the Phobious program within the Geneious plugin was used to predict functional domains in the amino acid sequences.
Confirmation of mutant phenotypes by allele swap.The mutations were further confirmed with Sanger sequencing and by swapping the mutant allele to the wild-type background and vice versa. Allele swap experiments of the selected incomplete resistant (IR4) and complete resistant mutants (R1, R3, and R4) were performed as previously described in Hockett et al. (31). Briefly, an ∼1-kb fragment containing the mutant or a WT allele was amplified with a Phusion high-fidelity DNA polymerase (New England BioLabs) using standard protocols. Primers used for generating the PCR fragments and mutant confirmation are listed in Table S2 in the supplemental material. The PCR fragment contained gateway cloning sites added through the primer extension. The purified PCR fragment was cloned into pDONR207 and further recombined into pMTN1907 using BP and LR clonase enzymes, respectively. pMTN1907 containing a desired clone was transformed to S17-1 and conjugated to the Pph wild-type or mutant background by biparental mating. Tetr merodiploids of Pph selected on Tet, Rif, and NFT plates were counterselected in KB supplemented with 10% sucrose, followed by PCR confirmation of the desired allele-swapped strains. Allele swap of the high persistent-like (HPL) mutant was performed similarly but used a one-step gateway vector, pDONR1K18ms, and the merodiploids were selected in Km, Rif, and NFT plates.
LPS extraction and visualization.LPS extraction was performed as described by Davis and Goldberg (79) from pellets of 1 milliliter of overnight cultures (optical density at 600 nm [OD600], 0.5). After extraction, 10-μl samples were separated by SDS-PAGE. Samples were labeled using the Pro-Q Emerald 300 glycoprotein stain kit (number P21857; ThermoFisher Scientific) according to the manufacturer’s instructions. For LPS size determination, CandyCane glycoprotein molecular weight standards (number C21852; ThermoFisher Scientific) were included. Gels were visualized using the Molecular Imager Gel-Doc XR+ system (Bio-Rad) with Image Lab software.
In planta fitness test of the high persistent-like, incomplete resistant, and resistant mutants.The high persistent-like (HPL), selected incomplete resistant, and complete resistant mutants of Pph, including a type III secretion mutant, ΔhrpL::Pph (80), were tested for their in planta fitness. An in planta experiment was performed in a growth chamber (Conviron) maintained at 24°C, 75% relative humidity (RH), and 16 and 8 h of day/night cycles. Plants of the dwarf French bean (Phaseolus vulgaris) variety ‘Canadian Wonder’ were grown in Dillen 6.0 standard pots (Onliant) in Sunshine mix 4 aggregate plus professional growing mix (Growerhouse). Plants were irrigated daily. Nine days postseeding, plants were infiltrated with a suspension of the bacteria. For inoculant preparation, overnight cultures of WT Pph, ΔhrpL::*Pph, and the tailocin persistent and resistant lines were pelleted by centrifugation, washed twice, and resuspended in equal volumes of 10 mM MgCl2 buffer. Optical density (OD600) was adjusted to 0.1 using a Spectronic 200 spectrophotometer (Thermo Scientific) and diluted 50 times. A total of ∼200-μl diluted cultures were infiltrated onto designated areas of the two primary leaves using 1-ml BD syringes, and infiltrated areas were marked. Infiltrated areas were harvested using a 1-cm cork borer in a 2-ml tube containing 200 μl 10 mM MgCl2 and 2 of the 3-mm glass beads (VWR) at 0, 24, and 48 h postinfiltration. Harvested leaf discs were homogenized in a FastPrep-24 instrument (MP Biomedicals) for 20 s. Homogenates (5 μl) after serial dilutions were spotted on KB agar plates supplemented with 50 μg/ml of nalidixic acid, and CFU were counted after 2 days. In planta experiments were repeated at least twice with 8 replications per time.
Statistical analysis.Means of total and surviving populations between treatments were compared using the Glimmix protocol in SAS 9.4 with experimental repeat used as a random factor. Whenever required, post hoc analysis was performed with Tukey’s honestly significant difference (Tukey HSD) test at a 5% significance level (P = 0.05). In planta enumeration data were analyzed in JMP Pro 14 (SAS Inc.) using Fit Y by X model and one-way ANOVA and Tukey HSD test at P = 0.05 after confirming that data were normally distributed and had equal variances.
Data availability.The read files of all the genomes generated in this study have been deposited to the NCBI Sequence Read Archive (SRA) database under BioProject PRJNA608702 and BioSamples SAMN14206700 to SAMN14206710. The SRA accession numbers are SRR11179092 to SRR11179102.
ACKNOWLEDGMENTS
This research was supported by the following sources for K.L.H.: the USDA National Institute of Food and Federal Appropriations under project PEN04648 and accession 1015871, the USDA NIFA Foundational Program (2019-67013-29353), and the Lloyd Huck Early Career Professorship.
We thank Emily Weinert from the Department of Biochemistry and Molecular Biology at Penn State University for her help with in silico protein work and Brian Kvitko from University of Georgia for gifting the plasmid vectors.
FOOTNOTES
- Received 16 March 2020.
- Accepted 16 April 2020.
- Accepted manuscript posted online 20 April 2020.
Supplemental material is available online only.
- Copyright © 2020 American Society for Microbiology.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
