Evidence of Tailocin Persistence and Resistance in Pseudomonas

Tailocins are bacterially-produced, phage tail-like bacteriocins that have been proposed as therapeutics for treatment of bacterial infections. However, we have a limited understanding of how target populations survive tailocin exposure. In this paper, we demonstrate that cells of a target population of Pseudomonas syringae are able to survive lethal doses of a tailocin through both physiological and genetic mechanisms with stationary phase cells predominantly surviving using a physiological mechanism. Regardless of growth phase, a significant fraction of cells that survived tailocin exposure did not exhibit any increased resistance to subsequent exposure, indicating that these cells survive by persistence rather than resistance mechanism. Of those cells that did gain a detectable increase in tailocin resistance, there was a range from insensitive (complete resistance) to partially sensitive. We also recovered a mutant exhibiting a high-persistence like phenotype that showed significantly increased survival to transient exposure but no detectable growth after prolonged tailocin treatment. By sequencing the genome of multiple types of mutants, we identified several genes linked to lipopolysaccharide (LPS) biogenesis and implicated in tailocin tolerance and resistance. In particular, we found a specific genomic region that, when mutated, gave rise to the various classes of resistance and the high persistence phenotypes. Furthermore, we showed that a hypothetical gene involved in the high persistence phenotype is transcriptionally fused with the LPS biosynthetic region and contains a signal peptide and several trans-membrane domains. While the complete resistant mutants had lost their LPS O-antigen, incomplete resistant mutants contained minor to significant changes in their O-antigen. This work demonstrates that Gram-negative bacteria can survive tailocin exposure through multiple strategies, including the first description of a persistence-like mechanism for tolerating tailocin exposure. Author Summary With the rise in antibiotic resistant infections, it has become a necessity to study and find alternative treatment strategies. Bacteriocins, which are bacterially produced protein toxins, have been proposed as one of such alternatives. However, a deeper understanding of how target bacteria respond to bacteriocin exposure is lacking. Here, using a phage-tail like bacteriocin known as a tailocin, we show that target cells of the important plant pathogen Pseudomonas syringae are able to survive tailocin exposure through both physiological persistence and genetic resistance mechanisms. We show that bacterial cells that are not growing rapidly rely primarily on persistence mechanism, whereas rapidly growing cells are more likely to survive through genetic resistance. By genomic sequencing, we identified various types of mutations in the genes involved in lipopolysaccharide (LPS) biogenesis that cause increased tailocin persistence and resistance. In particular, we identified a mutation in a LPS-related genomic region that encodes a hypothetical protein and causes tailocin high persistence. We conclude that tailocin persistence is a real phenomenon and effects of both persistence and resistance in long term disease management needs to be considered before designing tailocins into biocontrol agents.

both physiological and genetic mechanisms with stationary phase cells predominantly surviving 23 using a physiological mechanism. Regardless of growth phase, a significant fraction of cells that 24 survived tailocin exposure did not exhibit any increased resistance to subsequent exposure, 25 indicating that these cells survive by persistence rather than resistance mechanism. Of those cells 26 that did gain a detectable increase in tailocin resistance, there was a range from insensitive 27 (complete resistance) to partially sensitive. We also recovered a mutant exhibiting a high-28 persistence like phenotype that showed significantly increased survival to transient exposure but 29 no detectable growth after prolonged tailocin treatment. By sequencing the genome of multiple 30 types of mutants, we identified several genes linked to lipopolysaccharide (LPS) biogenesis and 31 implicated in tailocin tolerance and resistance. In particular, we found a specific genomic region 32 that, when mutated, gave rise to the various classes of resistance and the high persistence 33 phenotypes. Furthermore, we showed that a hypothetical gene involved in the high persistence 34 phenotype is transcriptionally fused with the LPS biosynthetic region and contains a signal 35 peptide and several trans-membrane domains. While the complete resistant mutants had lost their 36 LPS O-antigen, incomplete resistant mutants contained minor to significant changes in their O-37 antigen. This work demonstrates that Gram-negative bacteria can survive tailocin exposure 38 through multiple strategies, including the first description of a persistence-like mechanism for 39 tolerating tailocin exposure. 40

Introduction 56
The decreasing efficacy of commonly used antibiotics in treating bacterial infections of humans, 57 animals, and crop plants is a significant concern for plant and animal health. According to a 2013 58 report of Centers for Disease Control and Prevention (CDC), in the United States alone, 59 antibiotic resistant infections are estimated to cause two million infections and 23,000 deaths 60 annually (1), with the number of deaths predicted to reach 10 million per year by 2050 (2). 61 Although emergence of genetic resistance to antibiotics is a main driver of this trend, increasing 62 evidence suggests that non-heritable physiological persistence also plays a critical role in both 63 antibiotic treatment failures and infection relapse (3-6). Resistance is a result of genetic changes 64 that enable cells to withstand and grow in the presence of higher doses of an antibacterial agent, 65 whereas persistence is an altered physiological response of a sub-population of genetically 66 sensitive cells to survive transient lethal treatments (7). In addition to the altered physiological 67 response that is mainly governed by gene expression changes, persistence can be caused by 68 genetic mutations in genes associated to these physiological responses. Although resistance and 69 persistence are considered distinct survival strategies, recent evidence suggests that physiological 70 persistence can also reduce antibiotic efficacy by promoting the acquisition of resistance (8,9). 71 As such, finding alternatives to treat bacterial infections is critical to combat the reduced efficacy 72 of antibiotic therapies. More importantly, employing the alternative treatments in a manner that 73 sustains their efficacy is critical for durable pathogen control. Bacteriocins, which are bacterially 74 produced proteinaceous toxins, have been proposed as antibiotic alternatives (10). However, we 75 currently lack a fundamental understanding of how bacterial populations are likely to survive 76 bacteriocin treatments, should they be widely adopted. 77 Bacteriocins encompass diverse bacterial peptides, proteins, or protein complexes that are 78 antagonistic mostly toward phylogenetically close relatives of the producer (11, 12). This 79 potential of bacteriocins to inhibit the growth of competitor bacteria has been already utilized in 80 food preservation to reduce the growth of food spoiling and pathogenic bacteria (13). 81 Bacteriocins are classified into different groups based on their structure, composition, and mode 82 of action. Tailocins are bacteriocins that resemble bacteriophage tails and are grouped into R-83 type (with a retractile core tube) and F-type (flexible) (14, 15). Bacteria including the 84 opportunistic human pathogen P. aeruginosa and other environmental Pseudomonads and 85 Burkholderia are known to produce tailocins that are believed to antagonize competitors 86 including pathogenic strains among others (16-22). In fact, a recent study by Principe et al. 87 showed the effectiveness of foliar sprays of tailocins produced by P. fluorescens in reducing the 88 severity and incidence of bacterial-spot disease in tomato caused by Xanthomonas vesicatoria 89 (22). Other studies have also indicated the potential use of engineered R-tailocins in suppressing 90 food borne pathogens using in vivo models (23, 24). Our group has previously characterized a R-91 type tailocin from a plant pathogenic bacterium P. syringae pv. syringae (Psy) B728a (25). This 92 R-type tailocin showed antagonistic potential against several pathovars of P. syringae that cause 93 serious diseases and substantial losses in economically important crops such as common bean 94 (pv. phasiolicola), soybean (pv. glycinae), chestnut (pv. aesculi), and kiwifruit (pv. actinidae) 95 (25). A broad spectrum of tailocin mediated antagonistic interaction has been recently described 96 in P. syringae (26). 97 In these systems, tailocins are considered to be potent killers as a single tailocin particle is 98 predicted to kill a sensitive cell and an induced cell can release as many as 200 particles (27, 28). 99 R-tailocins kill their targets by binding to the surface receptors and puncturing through the cell 100 membrane causing dissipation of membrane potential and cell death (11,27 Major limitations related to employing bacteriocins stem from our lack of understanding of how 112 target pathogens might persist in the face of sustained treatment. Moreover, a detailed 113 understanding of the genetics of LPS modification in relation to the evolution of tailocin 114 resistance and persistence in the target pathogen is lacking. This knowledge is fundamental to 115 assessing the sustainability of tailocins as therapeutic agents. 116 In this study, we aimed to determine the tolerance and resistance responses of target cells 117 [Pseudomonas syringae pv. phaseolicola (Pph)] against tailocin produced by Psy. We exposed 118 stationary and logarithmic cultures of Pph to lethal doses of tailocin and screened these surviving 119 cells for resistant and tolerant phenotypes by tailocin re-treatment. We found that pronounced 120 differences between the frequencies of resistant and persistent cells depending on growth phase 121 of the target cells. Furthermore, we sequenced and analyzed the genomes of fourteen mutant 122 lines exhibiting different levels of resistance and tolerance to identify mutations potentially 123 contributing to each of these phenotypes. Of the heritable mutants, we identified ten unique 124 mutant alleles with likely roles in the lipopolysaccharide (LPS) O-antigen biosynthesis leading to 125 various degrees of tailocin resistance. We also identified an open reading frame (ORF) that is 126 present within the operon containing LPS biogenesis genes and which encodes a hypothetical 127 protein containing several transmembrane-domains and a signal peptide at its N-terminus. A 128 mutation within this hypothetical protein led to increased tailocin persistence and stationary and 129 log phase cells were no more different in tailocin tolerance for this mutant. 130

Materials and Method 132
Bacterial strains, media, and culture conditions 134 All bacterial strains and mutants used in this study are listed in Table 1. P. syringae pv. syringae 135 (Psy) wild-type (WT) strain B728a and its tailocin defective mutant ∆Rrbp were used to prepare 136 the treatment supernatants. P. syringae pv. phaseolicola (Pph) 1448A was used as the target 137 strain. Tailocin resistant and tolerant mutants of Pph generated in this study are described in 138

Tailocin induction and purification 142
Tailocin or control supernatants for treatment were prepared from logarithmic (log) cultures of 143 Psy B728a and ∆Rrbp, respectively using a polyethylene glycol (PEG) precipitation protocol as 144 previously described (25, 36). Briefly, 100-fold diluted overnight B728a cultures were sub-145 cultured for 4-5 hours in KB before inducing with 0.5 µg ml -1 final concentration of mitomycin C 146 (GoldBio). Induced cultures were incubated for 24 hours with shaking at 28 • C. Next, cells were 147 pelleted by centrifugation and the supernatants were mixed with 10% (w/v) PEG 8000 148 (FisherScientific) and 1M NaCl. Supernatants were then incubated either in ice for 1 hour and 149 centrifuged at 16,000 g for 30 min at 4 • C or incubated overnight at 4 • C and centrifuged at 7000g The relative activity of the purified tailocin was determined to be 10 3 -10 4 AU and 1.25×10 7 -160 4.25×10 9 lethal killing units/ml. The minimum inhibitory concentration was estimated to be 100 161 AU when exposed to ~10 6 viable target cells at their logarithmic growth. No loss of tailocin 162 activity was observed for a period of over six months in the buffer (10 mM Tris PH 7.0, and 10 163 mM MgSO4) at 4 • C. 164

Tailocin treatment and survival assessment for stationary and log cultures 165
To assess tailocin activity against the stationary and log phase of Pph, individual colonies 166 growing on KB agar plates for ~2 days were inoculated into 2 ml of liquid KB medium. 167 Following incubation at 28 • C with shaking at 200 rpm overnight, the cultures were back diluted 168 1000-fold into fresh KB. The back diluted cultures were either incubated for 28-30 hours to 169 prepare stationary cultures, or back-diluted 100-fold at 24 hours and cultured for another 4-6 170 hours to prepare log cultures (see Fig S1 for a growth curve of Pph). 171 Stationary cultures were diluted 20,000fold and logarithmic cultures were diluted 1,000-fold 172 [~10 5 -10 6 CFUs/ml for both cultures see Fig 1] in fresh KB before tailocin treatment. Treatment 173 was applied by mixing 10 µl of diluted cultures in 90 µl of purified tailocin preparation diluted in 9 KB. After treatment, samples were incubated for ~ 1 hour at 28 • C and washed twice to remove 175 residual tailocin particles. Washing was performed by mixing the treated culture in 900 µl of 176 fresh KB followed by centrifugation at 12,000 g for 2 min. The top 900 µl fraction was discarded 177 and the bottom 100 µl fraction was serially diluted and either spread-or spot-plated to enumerate 178 surviving population. Plates were incubated at 28 • C for 2-3 days before enumeration. Serial 179 dilutions of both stationary and log cultures were spotted onto KB agar to enumerate the 180 untreated population. Experiments were performed with various tailocin concentrations (i.e. 100 181 AU, 500 AU, and 900 AU). 182 Tailocin re-treatment to differentiate resistance and persistence 183 Surviving colonies were treated again with tailocin to differentiate them into resistant or 184 persistent colonies. Re-treatment was performed by an overlay method as described previously 185 (36), or by broth treatment as discussed above. Overlay method was used to determine the AU of 186 the tailocin preparation with the selected mutant lines. Broth exposure was used to calculate 187 reduction in the population of log cultures. Surviving colonies were differentiated into various 188 phenotypes as follows: persistent (sensitive to tailocin to the wild-type level in both the broth and 189 overlay method), high -persistent like (completely sensitive in the overlay but survived 190 significantly more than wild-type under broth condition), incomplete resistant (showed some 191 level of sensitivity at least in one of the treatment conditions), and complete resistant (were 192 insensitive under both conditions). 193

Time dependent death curve with tailocin treatment 194
Prolonged tailocin exposure was performed with both the stationary and log-phase cultures of 195 populations were enumerated at 1, 4, 8, and 24 hours following treatment and randomly selected 197 surviving colonies were re-exposed to tailocin to differentiate them into persistent or resistant 198 phenotypes. 199

Tailocin recovery from the treated samples and activity testing 200
Stationary and log cultures treated with tailocin for 1-24 hours as described above were 201 centrifuged and the supernatant was collected, and filter sterilized using a 0.22 µm syringe filter. 202 Supernatants were diluted 5-, 10-, 50-, and 100-fold in KB and spotted on Pph overlay. Purified 203 tailocin particles diluted in KB were also included as control treatment. 204

Determining the effect of stationary and log supernatant on tailocin activity 205
Stationary and log phase cultures were prepared as described above by culturing Pph cells in KB 206 broth for either 28-30 hours or for 4-6 hours, respectively. Cultures were centrifuged for 2 min at 207 12,000 g and the supernatant was filter sterilized using a 0.2 µm syringe filter. Stationary and log 208 supernatants were diluted 1,000-20,000-fold in KB (according to how the cultures were diluted 209 for tailocin treatment). Various dilutions (10-, 50-, 100-, and 1000-folds) of purified tailocin 210 were prepared in the stationary and log supernatants. Dilutions were spotted on a Pph lawn using 211 the overlay method. AU of the of the tailocin suspensions were calculated as described above. 212

LPS extraction and visualization 213
LPS extraction was performed as described by Davis and Goldberg (38) from pellets of one 214 milliliter of overnight cultures (OD600 0.5). After extraction, 10 µl samples were separated by 215 determination, CandyCane™ glycoprotein molecular weight standards (ThermoFisher Scientific, 218 #C21852) were included. Gels were visualized using Molecular Imager Gel-Doc XR+ (Bio-Rad) 219 with Image Lab Software. 220

Genome sequencing and analysis of the tailocin high-persistent and resistant mutants 221
The tailocin high-persistent like (HPL) and complete and incomplete resistant mutants recovered 222 from tailocin treatment of Pph wild-type cells ( Presence absence and homology searches of the selected genomic regions and genes implicated 244 in tailocin resistance and tolerance were performed with NCBI and IMG-JGI databases using 245 BLAST algorithm using the Pph sequences as query. InterProScan (41) and Phobious program 246 within the Geneious plugin was used to predict functional domains in the amino acid sequences. 247

Statistical analysis 248
Means of total and surviving population between treatments were compared using the Glimmix 249 protocol in SAS 9.4 with experimental repeat used as a random factor. Whenever required, post 250 hoc analysis was performed with Tukey's Honest Significant Difference (Tukey HSD) test at 5% 251 significance level (P=0.05). 252

Results 253
Tailocin persistence increased in the stationary state 254 Purified tailocin supernatant was used to test its killing effects on stationary and log phase 255 cultures of the Pph target cells in a broth environment. After an hour of 100 AU tailocin 256 treatment, a consistent reduction (3.59 ± 0.12 log reduction) in the viable population occurred for 257 logarithmic cultures, while a significantly lower reduction (1.38 ± 0.14 log reduction) occurred 258 for the stationary cultures. Further analysis showed that, upon treatment of equivalent number of 259 viable cells, stationary cells consistently survived 10 to 100-fold more than the logarithmic cells 13 (Fig 1 and Fig S2). Surviving colonies, especially those from the stationary phase, were 261 predominantly sensitive upon tailocin re-exposure suggesting survival by persistence mechanism 262 (see below). did not occur in either culture (Fig 2A). Twenty-four hours post-treatment, although the overall 270 population increased (Fig 2A), individual treatments showed different results: for some replicate 271 treatments the population remained constant suggesting maintenance of the persistence state, 272 while for some other replicates, population increased due to division of cells that acquired 273 tailocin resistance (see Fig S3). 274 Upon tailocin re-treatment, >90% of stationary and >60% of log cells that survived the first hour 275 treatment, were as sensitive as the WT (i.e. persistent) as in (Fig 2B). The proportion of 276 persistent survivors was higher in the stationary cultures compared to the log cultures at all time 277 points (Fig 2B). Tailocin persistent cells were recovered from both cultures even after 24 hours 278 of tailocin treatment, although the proportion decreased over time (Fig 2A). Tailocin activity 279 (see Fig S4) was detected in the supernatants recovered from the treated samples that contained 280 tolerant cells, confirming saturation of tailocin in the treatment. Although a slight reduction of 281 activity was observed when the tailocin preparation was mixed with undiluted stationary 282 supertant compared to log supertanant, no difference was detected upon diluting the supertants 283 up to 1,000-20,000-fold (as the cultures were diluted for tailocin treatment) before mixing with 284 tailocin (Fig S5). This suggested that the increased tailocin tolerance in the stationary phase is 285 not related to inhibition of tailocin activity by an extracellular component. 286 Upon treating the cells with a concentrated tailocin (900 AU) the surviving population decreased 287 such that no difference in survival between the stationary and log cultures was detected (Fig 3A). 288 However, even with this higher level of tailocin applied, the proportion of tailocin persistent cells 289 remained higher for stationary phase survivors than that for the log phase survivors (Fig 3B). 290

Tailocin exposure selected for heritable mutants showing heterogenous resistance and 291
increased persistence 292 In addition to the recovery of tailocin persistent sub-population, we recovered complete tailocin 293 resistant mutants (i.e. heritable mutants that are insensitive to lethal doses of tailocin), and 294 incomplete resistant (IR) mutants (see Fig. 2B and 3B for proportion) that were still sensitive to 295 tailocin, but with sensitivity significantly decreased compared to the wild-type cells both in 296 liquid-broth and agar-overlay conditions (Fig 4A and 4B). Moreover, we recovered an unique 297 phenotype, refered here as high persistent-like (HPL), that, compared to the wild-type, showed 298 increased persistence in liquid-broth treatment but equivalent sensitivity in an agar overlay 299 setting (Fig 4A and 4B). Both the IR and HPL phenotypes were heritable as progeny colonies 300 carried the same phenotype. Furthermore, the HPL phenotype did not differ in survival between 301 the stationary and log phase (Fig 5). 302 LPS analysis of the mutants and the wild-type Pph showed that the complete resistant mutants 303 lacked fully-formed O-antigens, whereas the high-persistent and majority of incomplete resistant mutants still possessed the O-antigen with subtle changes. One of the incomplete resistant 305 mutants (IR4), however, showed a very different and faint O-antigen band (Fig S6).  Table 2). Mutants isolated at different experiments showed mutation in a 315 different gene. A specific region (Fig 6) was identified in the Pph genome that showed the most 316 prominent role in tailocin tolerance and resistance. Even within a single gene (PSPPH_0957), 317 different phenotypes were identified depending on the nature of the mutation (see Table 2 and 318

Bioinformatics of mutated genes 324
Genes identified to have a role in tailocin activity (see Table 2 for a list) were assessed for their 325 presence and similarity to those found in other Pseudomonas genomes. PSPPH_0957 is 326 predicted to encode a glycosyl transferase (family 1) and conferred both complete and 327 incomplete resistance phenotypes depending on the type of mutation (see Fig 6) and is present 328 across majority of sequenced P. syringae isolates. However, the region from PSPPH_0958 to 329 PSPPH_0964 displays much more variation within P. syringae. For instance, a pv. glycinea race 330 4 isolate, which is closely related to Pph, lacks this region. Although predicted orthologs for 331 some of the genes in this region were present, PSPPH_0963, and PSPPH_0964 that occur 332 contigiously in an operon were less abundant even within P. syringae (present only in 87 333 genomes out of 287 P. syringae genomes in IMG-JGI database). Orthologs of these two genes 334 were not found in the genomes of pv. tomato (DC3000), Psy (B728a), and Pgy (R4), but were 335 found in strains of pv. actinidiae, pv. morsprunorum, pv. aesculi and some environmental 336 isolates. Although the incomplete resistance gene (PSPPH_0963) was identified to have a 337 FAD/NAD(P) binding cytoplasmic domain with a potential role in LPS biogenesis (IMG 338 product name UDP-galactopyranose mutase), no functional role could be predicted for the 339 hypothetical gene involved in the high-persistent like phenotype. Nevertheless, it possesses ten 340 predicted transmembrane domains and a signal peptide domain at its N-terminus. 341

Discussion 342
There has been a renewed research interests in alternative treatment strategies for bacterial 343 pathogens due mainly to the growing threats of antibiotic resistant infections. Tailocins have 344 long been proposed as effective and more specific alternatives to broad spectrum antibiotics. 345 However, the evolutionary dynamics of tailocin persistence and resistance have not been widely 346 investigated. In this study, we addressed this fundamental concern using a phage-tail like bacteriocin (i.e. tailocin) produced by P. syringae pv. syringae strain B728a in killing target cells 348 of P. syringae pv. phaseolicola 1448A. Upon expoure of a lethal dose of tailocin to equal 349 numbers of stationary and logarithmic Pph cells, a higher fraction of stationary cells survived 350 than cells in log phase. Upon re-exposing surviving cells, a majority of the cells were sensitive to 351 the wild-type level, suggesting that non-heritable persistence rather than heritable resistance was 352 the predominant survival mechanism , particularly in the stationary phase. A prolonged tailocin 353 exposure generated a killing pattern similar to the one reported for persistent sub-population 354 upon antibiotic treatment (7, 43). Persistence was maintained for at least 24 hours with tailocin 355 exposure, a phenomenon that was more evident in some experimental replicates in which 356 resistance evolution was not observed. However, by increasing the tailocin concentration in the 357 treatment, we showed that most of the persistent survivors were killed, and the difference in 358 survival between the two growth phases was no longer seen. Upon re-exposure, however, 359 stationary phase-derived cells exhibited a higher fraction of persistence. Overall, these 360 experiments using various tailocin concentration and exposure time suggested that tailocin 361 persistence is dependent on concentration rather than duration of exposure. We showed that 362 increasing the concentration of the tailocin treatment enabled killing of even the highly persistent 363 stationary cells. This indicated that the stationary cells may require multiple hits by tailocin 364 particles as opposed to the one-hit-one-kill mechanism of killing described for tailocins (27, 28), 365 or that the probabilty of a successful hit in stationary phase is lower than in log phase. Since 366 tailocins are specific to their targets, and are not known to have off-target effects, higher 367 concentration of tailocin could be used to achieve a more effective pathogen control. However, 368 although at a low level, persistence was still maintained even with high-dose tailocin treatment 369 and inherent emergence of either complete or incomplete resistance was frequently observed. As 370 such, although a significant reduction in pathogen population and disease pressure can be 371 obtained with tailocins, a stand-alone tailocin treatment might not be enough to achieve a 372 sustainable pathogen control. 373 The use of the term 'persistence' in relation to antimicrobial survival is disputed to some extent 374 and is sometimes used interchangably with 'tolerence' and 'viable but not culturable state'. In 375 this paper, we used the term 'persistence' as this trait was only seen in a sub-population, resulted 376 in a bi-phasic death curve, the surviving persister cells resuscitated in the absence of the tailocin 377 treatment, and were equally sensitive to the wild-type cells upon re-exposure. This definition of 378 persistence has been suggested previously (7). Persistence to antimicrobials is being increasingly 379 persister-subpopulation (4,6,48). TA systems were shown to be induced when cells were 385 starved for certain sugars and amino acids or by exposure to osmotic stresses that altered ATP 386 levels in the cell (47, 49). However, TA system activation did not always induce persister 387 formation (49). Additionally, recent findings have indicated a mechanism mediated by the 388 guanosine pentaphosphate/tetraphosphate (ppGpp) for persister formation that is not dependent 389 on a TA system (50). Instead, a strong stationary state effect, that likely involved starvation 390 response, was shown to increase persistence by 100-1,000 fold in Staphylococcus aureus with 391 ciprofloxacin treatment (49). Whether similar mechanisms of TA and/or ppGpp systems regulate 392 tailocin persistence or a specific mechanism for tailocin and/or related bacteriophage exists, 393 remains to be determined. Nevertheless, our data of the difference in tailocin tolerance between 394 the stationary and log cultures suggests that metabolic inactivity and starvation induced stress 395 could be a strong factor in tailocin persistence. Another mechanism that can cause increased survival to surface active antimicrobials (eg. 411 phages and host immune defenses) is phase variation (54). Phase variation is a gene regulation 412 system that induces heterogenous expression of specific genes in a clonal population (54-56). 413 Phase variation is heritable but reversible as the cells go through generations. The 'ON' 'OFF' 414 switch occurs randomly amounting to 10 -4 to 10 -1 per generation, significantly greater than is 415 expected by mutational events (57). Phase variation has been shown to modify LPS operons in S. 416 enterica spp (58). Moreover, temporal development of phage resistance in S. enterica serovar 417 Typhimurium by phase variable glucosylation of the O-antigen has been evidenced (59). In 418 addition, phase variation in P. fluorescence has been demonstrated during rhizoshpere 419 colonization of Alfala. The phase variants were enhanced in motility by production of a longer 420 flagella than the wild type cells and were able to colonoze distal end of the roots (60). It can be 421 expected that phase variation could be a part of tailocin survival mechanism, particularly because 422 of the fact that LPS, that is commonly modified in bateria by phase variation, serves as the will be required to determine the rate of switch to confirm if this is a case of phase variation. 432 To our knowledge, the incomplete resistant phenotypes observed here have not been described 433 before with bacteriocins although few studies have discussed bacteriocin persistence (66, 67) . 434 Here, using a phage-tail like bacteriocin, we showed that resistant lines with various degrees of 435 sensitivity are selceted by exposure of target cells to tailocins. As the complete phage resistant 436 mutants have been shown to be defective in virulence and fitness (68), incomplete resistance, 437 which likely has a less pronounced fitness cost, might be a better survival strategy. In the 438 incomplete resistant mutants, the LPS O-antigen region was present, but modified at a gross 439 level, unlike in the complete resistant mutants that had completely lost their O-antigen. 440 Preserving the O-antigen and still surviving the tailocin or phage attack may prevent the fitness 441 cost associated with complete resistance evolution. Further tests that will compare the virulence 442 and fitness traits of the incomplete resistant mutants together with the wild-type and resistant 443 lines will be required to confirm these hypotheses. 444 The LPS region of Pph containing genes involved in incomplete resistance and tolerance was not