Systematic functional assessment of anti-phage systems in their native host

35 The balance between resistance and susceptibility of bacteria to their viruses (bacteriophages, 36 or phages) is governed by two primary strategies. The first consists in preventing the 37 attachment of phages to the bacterial cell surface. This can occur through the loss, 38 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 2 modification, or masking of phage receptors, effectively rendering the host inaccessible to 39 infection. The second blocks post-adsorption steps in the phage lifecycle, a process mediated 40 by specialized defence systems. In recent years, more than a hundred novel defence systems 41 have been identified, revealing a remarkable diversity of mechanisms. These systems include 42 those that degrade invading phage DNA, such as Restriction-Modification (RM) systems and 43 CRISPR-Cas systems. Others inhibit phage gene expression or replication, exemplified by 44 mechanisms like Viperins and chemical -based defences (1–3). Additionally, some defen ce 45 systems rely on extreme strategies, including triggering programmed cell death or inducing 46 dormancy, to stop phage proliferation. These mechanisms are known as abortive infection 47 systems (4). The variety of such defen ces reflects the constant and dynamic evolutionary 48 struggle between bacteria and their viral predators, as highlighted in recent literature (5). 49 Computational analyses based on remote sequence homology searches have predicted that 50 the bacterial defensome (i.e., the set of defence genes present in the same genome) is 51 composed of 5-6 defence systems in average, although this number is highly variable, even 52 between related strain (6). These observations have given rise to the view of this defensome 53 as an integrated bacterial immune system, in which individual defence systems are involved 54 in a complex network of interactions. Supporting this view, synergistic associations of defence 55 systems have been reported recently (7–12). Carrying multiple defence systems in a single 56 genome is thought to have two main positive effects on bacterial fitness: it protects against a 57 larger diversity of phages (11,13), as individual defence systems may have a narrow specificity 58 and restrict only few phage genera (13), and it limits the emergence of phages that escape 59 bacterial defences (7,11,14,15). However, we still know very little about how defence systems 60 interact with one another , how they are regulated and how they impact the ecology and 61 evolution of phage-bacteria interactions. One reason for this knowledge gap is that research 62 necessarily needed an initial effort focused on scaling up the discovery and characterisation 63 of defence systems, which rely on novel informatic tools and synthetic biology. Consequently, 64 most defence systems were studied from their heterologous expression in model organisms 65 such as Escherichia coli and Bacillus subtilis (16–18) . Comparatively, fewer defence systems 66 have been studied in their original host, which hinders our understanding of the activity, 67 regulation, fitness costs and benefits of defence genes (and their combinations) in their native 68 condition. Recent studies have interrogated whether and how the composition of the 69 defensome impacts the range of phages capable of infecting a host. Intuitively, one might 70 expect a negative correlation between the number of phages infecting a host and the number 71 and diversity of host defences . However, experimental findings have yielded contrasting 72 results. A recent study evaluating interactions between clinical isolates of Pseudomonas 73 aeruginosa and its phages found a positive correlation between the number of defense 74 systems encoded in the isolates and their resistance to the tested phages (13). In contrast, a 75 large-scale screening of phage-bacteria interactions in Escherichia coli found that, within this 76 cohort, phage resistance was more likely determined by adsorption factors whereas a very 77 low correlation between the number of defence systems and the level of phage susceptibility 78 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 3 could be observed (19). This suggests that, in this dataset, defen ce systems played only a 79 marginal role in shaping the range of phages capable of infecting a host. Additionally, this 80 study observed that the host range of specialist and generalist phages tended to overlap, 81 generating an interaction matrix with a nested pattern. 82 In another study, the investigation of interactions between natural , longitudinally sampled, 83 populations of Vibrio and their phages generated a modular interaction matrix: a specific 84 subset of phages infects a specific subset of related hosts and thus, the host range of phages 85 from two distinct subsets display very little overlap (20). In this case, the lack of interactions 86 between phages and bacteria that belong to distinct modules is due to lack of adsorption . 87 Conversely, the lack of interaction between phage s and bacteria that belong to the s ame 88 module is not explained by adsorption factors but correlate s with the presence of defence 89 systems (20). In sum , the contribution of the defensome in determining phage -bacteria 90 ecological interactions remain s unclear. In addition, the functionality and role of defence 91 systems in their native host have been overlooked, and therefore, little is known about their 92 levels of expression, their regulation and their interactions with other defence systems. 93 In this study, we aimed to complement ecological interaction matrices by providing a genetic 94 approach to assess the contribution of defen ce systems in shaping phage host range. 95 Specifically, we employed a systematic genetic strategy to directly evaluate the activity of 96 predicted defence systems in their native bacterial hosts. This approach allowed us to quantify 97 the respective contributions of individual defen ce systems in restricting the host range of 98 phages or modulating their infectivity levels. By integrating these direct genetic assessments 99 with ecological interaction data, we sought to bridge the gap between theoretical predictions 100 and empirical observations, offering a more nuanced understanding of the interplay between 101 phage resistance mechanisms and host susceptibility. 102 103 104

105 106 Phage sensitivity of E. coli clinical isolate NILS69 is mainly explained by adsorption factors 107 108 To test the role of defence systems in shaping phage susceptibility, we chose the clinical E. 109 coli B2 phylogroup, uropathogenic isolate NILS69 as a model system. This strain is part of a 110 collection of E. coli Natural Isolates with Low Subculture passages (NILS), developed to 111 provide the community with E. coli natural isolates that cover the diversity of phylogroups 112 while remaining as close as possible to the original isolated strain (21). 113 We first assess ed the ability of phages to infect wildtype (WT) NILS69 by screening the 114 recently published Antonina Guelin (AG) collection, which gathers virulent dsDNA phages 115 belonging to 19 different genera (19), through solid and liquid infection succeptibility assays 116 (Supplementary Figure 1 A,B). Ten phages out of 93 tested were able to affect the growth of 117 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 4 NILS69 in solid infection assay while only 8 affected NILS69 growth in liquid (Supplementary 118 Table 1). 119 To determine whether the infectivity of certain phages could be revealed under different 120 culture conditions, we varied the parameters of the solid infection assay, including incubation 121 temperature (30°C, 42°C) and salt concentration (by adding 30 mM MgSO₄ and 15 mM CaCl₂ 122 to the growth medium), but none of these conditions changed the range of infecting phages 123 (Supplementary Figure 1A). 124 125 We next characterized the infectivity of these 10 phages in more details . In solid medium, 126 they form plaques with low efficiency in efficiency of plating (EOP) assays (EOP ranging from 127 10-5 to 10-1), except for phage NIC06_P2 ( Figure 1A ). When introduced at a multiplicity of 128 infection (MOI) of 1 in liquid medium, they show different impact on bacterial growth, which 129 is either unaffected (536_P1, P6, P7, P9), delayed (LF73_P4, DIJ07_P1, AN24_P4) or strongly 130 suppressed (LF73_P1, LF73_P3, NIC06_P2) ( Figure 1B ). The effect of phages on bacterial 131 growth are consistent in liquid and solid assays: the phages with highest EOP cause the 132 strongest reduction of bacterial growth in liquid, whereas phages with low EOP do not affect 133 bacterial growth. The only exception is phage LF73_P4, which ha s little effect in liquid 134 infection assay despite having a relatively high EOP . 135 136 Next, time-resolved adsorption assays were performed for each of the 93 phages on NILS69 137 to test whether the host range defined by our initial screening was due to a defect in phage 138 adsorption. We confirmed that the 83 phages defined as non-infective by our screen failed to 139 bind to NILS69 (Supplementary Figure 2 ). Therefore, the susceptibility of NILS69 to these 93 140 phages appears to be exclusively determined by adsorption factors. 141 142 Restriction Modification and PD -T4-3 reduce the virulence of infecting phages 143 144 We next wondered whether intracellular defences could impact the infectivity levels of the 10 145 phages that infect NILS69. NILS69 carries 6 defence systems as predicted by PADLOC (22) 146 and DefenseFinder (6) (Figure 1 C and Supplementary Table 2), as well as 7 additional 147 putative systems predicted by PADLOC. Individual deletion mutants of PD -T4-3, PD-T7-1, 148 PsyrTA, SoFIC and Thoeris were constructed (Figure 1D). Regarding the predicted type II RM 149 system, computational tools identified a HNH enconuclease associated with two 150 methyltransferases (red arrows in Figure 1E ). Manual inspection of the neighbouring region 151 revealed two additional HNH endonucleases (referred to as HNH endonuclease 1 and HNH 152 endonuclease 3, yellow arrows in Figure 1E) encoded in direct vicinity of the type II RM, which 153 were not detected by prediction algorithms. A BLAST search against the REBASE database 154 (23) indicated that HNH endonuclease 1 and HNH endonuclease 3 are similar to two putative 155 type IV endonucleases (Osp6506McrB2P and ObaORFAP with e -values of 4E-40 and 7E-20 156 respectively), although these were predicted based on structural prediction and not verified 157 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 5 experimentally. As these two endonucleases might have an effect on phages, we knocked out 158 a large 10.6-kb region, that we refer to as RM island, encompassing the 3 HNH endonucleases 159 as well as the two methyltransferase s. Out of the seven putative defence systems predicted 160 by PADLOC (22), we found PDC-SO4 and PDC-M63AB of particular interest as the first may 161 be similar to a Type II Toxin Antitoxin system (according to search against TADB3.0 (24)) and 162 the later consists in two genes that are encoded within a prophage, which are known to be 163 hotspots of defence systems (6,25–27). For these reasons, we included deletion mutants of 164 these 2 putative defence systems for downstream analyses. 165 The infectivity of the 10 phages was measured on each mutant through EOP and liquid assays 166 (Figure 1 D and Supplementary Figure 3 ). Increased EOP were observed for all phages, 167 except for NIC06_P2 and 536_P9, on mutants lacking RM island and PD-T4-3, indicating that 168 these defence systems protect NILS69 against these phages, while the other systems do not 169 seem to contribute. Because our previous screen showed difference between liquid and solid 170 conditions, we perform similar tests in liquid infection assays and found that while 536_P9 is 171 not affecting the growth of NILS69, this phage could effectively suppress the bacterial growth 172 of the RM island mutant, indicating that RM is also active against this phage (Supplementary 173 Figure 3). 174 We found that the RM island and PD -T4-3 target a non-overlapping set of phages, with the 175 former restricting 7 phages (536_P1, 536_P6, 536_P7, 536_P9, AN24_P4, DIJ07_P1, LF73_P4) 176 and the latter specifically inhibiting LF73_P1 and LF73_P3. To investigate potential negative 177 or positive interactions between defen ce systems, we generated four double mutants and 178 four triple mutants and tested their phage sensitivity using solid assays. Among these mutants, 179 only the double mutant lacking both PD-T4-3 and the RM island exhibited increased sensitivity 180 to phages, which was comparable to the additive effect observed in the corresponding single 181 mutants (Figure 1D ). Altogether, these data indicate that the RM island and PD-T4-3 are the 182 only active defence systems in NILS69 under these conditions and against this collection of 183 phages, and that their combined activity provides an additive protection against a broader 184 range of phages. 185 186 Experimental validation of the activity of an HNH endonuclease 187 Next, we aimed to identify which endonuclease within the RM island mediates phage 188 protection, hypothesizing that the type II RM system detected by DefenseFinder and PADLOC 189 (HNH endonuclease 2), which shows 100% identity to several restriction enzymes in REBASE, 190 is likely responsible for this phenotype. We constructed two mutants with deletions in different 191 portions of the RM island and measured phage EOP ( Figure 1E ). Only the mutants lacking 192 HNH endonuclease 3 (ΔRM_right) exhibited a phage sensitivity profile comparable to the 193 mutant lacking the entire RM island. These findings indicate that the type II endonuclease 194 detected by DefenseFinder and PADLOC is not responsible for phage protection. A sequence 195 homology search using BLASTP against the Standard database revealed that HNH 196 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 6 endonuclease 3 has 100% identity to an HNH endonuclease found in other E. coli isolates 197 (Uniprot: A0A3K0QCZ9). A structure -based homology search using AlphaFold3 (28) and 198 FoldSeek (29) identified a match to the Type IV methyl-directed restriction enzyme EcoKMcrA 199 encoded by E. coli K-12 MG1655 (E-value: 2.48E-04, run in Nov. 2024). Our data indicate that 200 the previously untested HNH endonuclease 3 has an antiphage activity and shares structural 201 features with a Type IV restriction enzyme (Figure 1E , Supplementary Figure 4 ). 202 Evaluation of the activity of NILS69 defence systems in the E. coli laboratory K-12 203 The apparent lack of activity of most of predicted defence systems might be linked to their 204 target specificity. In other words, they might be active against other phages than those we 205 tested. To explore this possibility, we tempted to cloned each individual defence system 206 under its native upstream regulatory sequence into a low copy plasmid and subsequently 207 transformed the constructs in E. coli K-12. Despite repeated attempts, we were unable to 208 clone Thoeris and the RM island. The K-12 strain has a distinct phage sensitivity profile 209 compared to NILS69, being sensitive to 15 of the 93 phages tested ( Supplementary Figure 210 1A). Comparing the infectivity of the 15 phages on K-12 and on clones complemented with 211 individual defence systems confirmed that PD-T4-3 is active and protects against 4 phages 212 (CLB_P2, LF73_P1, NIC06_P2 and T4 LD) which all belong to the Tevenvirinae subfamily 213 (Figure 1F ). Previous work indicated that PD -T4-3 target s phages from the Tequatrovirus 214 genus (T2, T4 and T6) (18). Our data suggest that the specificity of PD -T4-3 is not limited to 215 Tequatrovirus as it targets phage from Dhakavirus genus (CLB_P2) but PD-T4-3 likely does 216 not target all genera from the Tevenvirinae family since phages from Mosigvirus (LF82_P8) do 217 not seem affected. Interestingly, phage NIC06_P2, which is unaffected by PD-T4-3 in NILS69, 218 becomes sensitive to this defen ce system when overexpressed in K -12. This suggests that 219 NIC06_P2 may possess a counter-defence strategy that is no longer effective under this 220 condition of defence overexpression (Figure 1F ). This assay also revealed that PD -T7-1 221 protects against the phage PDP351_P2 ( Teseptimavirus) and has a partial protective activity 222 as it reduces the plaque size of phage T7. None of the other defence system had a detectable 223 protective activity in these conditions (Figure 1F). Surprisingly, the EOP of phages LM40_P1, 224 LM40_P2 and LM40_P3 were higher when the strain expressed PsyrTA. Because we also 225 noticed that this strain had a growth defect, we tested whehter the effect observed on LM40 226 phages was due to bacterial density. We show that phages LM40_P1, LM40_P2 and LM40_P3 227 have higher infectivity on low density bacterial lawns (Supplementary Figure 5) 228 229 Restriction modification system of NILS69 limits plasmid transmission 230 231 We reasoned that some of the defence systems encoded in NILS69 may target other type of 232 mobile genetic elements than phages. Therefore, we assessed their capacity to limit plasmid 233 transmission. We assessed the efficiency of transfer of a pRCS30, a 157kb - IncC plasmid 234 carrying multiple antibiotic resistance genes including an extended -spectrum b-lactamase 235 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 7 gene CTX-M14 (30), from its original host strain E. coli 513 to NILS69 and its derivative 236 mutants. 237 The conjugation efficiency, measured as the ratio of transconjugants to recipients, was 238 approximately 10⁻⁷ in NILS69, peaking at 1 hour. Similar efficiency was observed in all mutants 239 except for the ΔRM strain, which exhibited a 100 -fold increase (~10⁻⁵), suggesting that the 240 RM island restricts plasmid horizontal transmission. (Figure 2A ). To test which endonuclease 241 from the island was mediating this effect, we measured conjugation efficiency in the 242 DRM_right and DRM_left mutant backgrounds and found that the same HNH endonuclease 3 243 was responsible for protection against both phages and plasmids (Figure 2B). 244 Because this endonuclease share some similarity to a putative Type IV endonuclease, we 245 aimed to verify if it was indeed targeting methylated invading DNA or not. To test this, we 246 recovered transconjugants in NILS69, DRM_right and DRM_left backgrounds and measured 247 the efficiency of a secondary transfer of pRCS30 in NILS69 and DRM island mutants. Plasmids 248 from the donor strains NILS69 and DRM_left shoud be methylated, as these strains encode 249 the two methyltransferases in the RM island, while plasmids from the donor DRM_right strain 250 are expected to be unmethylated (Figure 1E). First, we observed that the transfer of pRCS30 251 into NILS69 is 1000 -fold higher when the plasmid ha d previously replicated in the same 252

compared to its original host E. coli 513 (Figure 2C). Second, we confirmed that 253 this increased conjugation efficiency is due to the presence of methyltransferases in the donor 254 strains, as their absence prevents this increase in transfer efficiency ( Figure 2C ). Our data 255 therefore indicate that NILS69 encodes an HNH endonuclease that restricts plasmid and 256 phage infection and which activity is likely inhibited by methylation. 257 Finally, we tested whether plasmid vertical transmission could be affected by the two defence 258 systems that are active in NILS69; PD-T4-3 and RM. To this aim we transformed the low copy 259 plasmid pZS into NILS69 WT, DRM or DPD-T4-3. Plasmid-carrying bacterial clones were initially 260 selected in the presence of antibiotics and subsequently transferred on non-selective media 261 to measure the percentage of plasmid loss in each background (Figure 2D ). While a fraction 262 of the population lost the plasmid, this did not depend of the bacterial genetic background, 263 therefore RM and PD-T4-3 had no effect on the vertical transmission of this plasmid. 264 265

266 267 The role of defence systems in shaping phage-bacteria interactions is still unclear. In particular, 268 the benefits provided by the accumulation of defence systems in a single genome remain an 269 open question. In this study we used a direct genetic approach to investigate the impact of 270 defence systems on the vulnerability of the E. coli clinical isolate NILS69 to a diverse collection 271 of phages. Confirming a previous study, we experimentally show that the inability of phage 272 to produce a successful infection is mainly explained by the lack of adsorption. Out of the six 273 predicted defence system s, only two have a measurable impact on the infectivity of the 274 phages that can bind the host; Restriction Modification and PD -T4-3. In addition, we tested 275 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 8 two putative systems (i.e., never previously verified experimentally) predicted by PADLOC 276 (22,31), namely PDC-M63AB and PDC-S04, but could not find evidence of defence activity in 277 our model system . These findings raise questions about the activity of other predicted 278 defence systems. We can foresee several reasons for the apparent inactivity of these systems. 279 First, the limited diversity of our phage collection, which consists entirely of virulent, double-280 stranded DNA phages, may introduce bias . Previous work has showed that certain defence 281 systems are specific and target a narrow range of phages (13,18) and it may just be that we 282 have not tested the right phages. Curiously, we showed that phage NIC06_P2 can be targeted 283 by PD-T4-3 when expressed in trans in E. coli K-12, but is not targeted in the native context 284 of NILS69. Therefore, a second possibility is that some of these defence systems may be 285 strongly regulated and expressed in different environmental conditions or may require 286 additional cofactor to act. Although some regulators have been identified (11,32,33), whether 287 defence systems are expressed and differentially regulated depending on the environment 288 remains an unexplored area. Future studies aiming at measuring the impact of the 289 environmental context on defence systems activity and their consequences on phage activity 290 will be key to advance our understanding on the ecological impact of defence systems. A 291 third explanation to the apparent inactivity of defence systems is given by phage -encoded 292 counter-defence proteins. Many counter-defences have been identified recently in phages or 293 plasmids and it clearly appears that many more remains to be uncovered (34). A recent effort 294 was made to catalogue and detect all known counter-defence proteins from genomes (35). In 295 this study, we observed that phage NIC06_P2, which has the highest EOP on NILS69 despite 296 an apparent low adsorption efficiency (Figure 1A, Supplementary Figure 2 ), is unaffected by 297 the defence systems tested since it lyses NILS69 as efficiently as all the derivative mutants. 298 Interestingly, this phage encodes a predicted anti-restriction nuclease (Arn) protein, which, in 299 phage T4, inhibits the modification-dependent endonuclease McrBC (Supplementary Table 300 1). Phage LF73_P4 also encodes a predicted anti -RM protein, Dam, which may partially 301 protect the phage against the RM island. 302 Interrogating the effect of defence systems on other mobile genetic elements, we found that 303 an HNH endonuclease effectively limits both phage and plasmid horizontal transmission. 304 Interestingly, this endonuclease shares structural similarity to a Type IV methylcytosine -305 targeting restriction enzyme but our data indicate that the methyltransferase that are encoded 306 directly downstream on the opposite strand limit the endonuclease activity, suggesting a Type 307 I,II or III mode of action. This association of an HNH endonuclease with two methyltransferases 308 might be conserved as our preliminary analysis detected this 3-gene association in E. coli, 309 Enterobacter bugandensis, Klebsiella pneumoniae and Vibrio alginolyticus. Further molecular 310 investigation of this endonuclease and associated methyltransferases will help to clarify their 311 mechanism of action. 312 In summary, we systematically assessed the functionality of defence systems encoded in a 313 clinical isolate. Our study indicates that, for this strain and in the conditions tested, the main 314 phage barrier is the bacterial envelope, while defence systems play a more marginal role and 315 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 9 further highlights the need to study defence systems in their native hosts, in variable 316 environmental conditions to improve our understanding of their roles in the ecology of phage-317 bacteria interactions. 318 319

and Method 320 321 Bacterial strains, phages, plasmids and g rowth condition d 322 Bacterial strains, phages, plasmids and primers used in this study are listed in Supplementary 323 Table 3. The medium used for all experiments was Lysogenic Broth (LB) prepared according 324 to the Luria formulation. Unless otherwise specified, bacterial cultures were inoculated in LB 325 from frozen stocks and propagated at 37°C with agitation. When needed, antibiotics were 326 added at the following concentrations: ampicillin 100 µg/m L, kanamycin 100 µg/m L, 327 chloramphenicol 30 µg/ml, streptomycin 30 µg/mL, and rifampicin 100 µg/m L. Solid or soft 328 media were prepared by adding 15 g/L or 7 g/L of agar, respectively. 329 330 Phage efficiency scoring 331 Phage efficiency of plating (EOP) against NILS69 and K12 -MG1655 derivative was tested by 332 spotting on soft agar plate 3µl of 10 -fold serially diluted phage suspensions. Soft agar was 333 inoculated to a final concentration of ≈10-7 CFU/mL from exponential phase growing cultures 334 and then incubated overnight at 37°C. EOP was calculated by dividing the Plaque Forming 335 Units (PFU) obtained on NILS69 by the number of PFU obtained on the phage production 336 strain. All the bacterial growth kinetic have been done in microtiter plate using “Tecan Infinite 337 M Nano” microplate reader. Wells were inoculated with 10 5 CFU/mL from overnight culture 338 with or without phages at a multiplicity of infection of 1. OD600nm readings were taken every 339 5 minutes over an 9h incubation period at 37°C with constant agitation 340 The screening of E. coli NILS69 against the Antonia Guelin phage collection was performed 341 in microtiter plate using 200 µL culture inoculated with 105CFU/mL from an overnight culture 342 and phages at a 100x dilution from stocks. OD readings were taken every 5 minutes over a 343 9h incubation period at 37°C with constant agitation. 344 345 Adsorption assays 346 Adsorption assays on NILS69 were conducted according to a previously described 347 protocol(36). Overnight bacterial cultures incubated at 37°C were transferred into a deep-well 348 plate. Phages were introduced at MOI 0.001 in a final volume of 1mL, with each well 349 representing a unique phage/bacteria combination. Adsorption was performed at 37°C for 350 30 minutes without agitation. To monitor adsorption kinetics, 150 µL samples were collected 351 at 10, 20, and 30 -minute intervals, treated with 50 µL of chloroform, and the supernatants 352 were titrated using a spot assay. Controls included a negative one with phages only (no 353 bacteria) and a positive one with phages and their reference bacterial strain, both of which 354 were also titrated. 355 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 10 356 Deletion mutants of NILS69 357 Deletion mutants of the defence systems in NILS69 were constructed using site -specific 358 recombination mediated by the l-Red system, as described by Datsenko and Wanner (37). 359 The primers listed in Supplementary Table 3 were used to amplify resistance cassettes via 360 PCR, generating linear DNA fragments containing homology regions for recombination. This 361 linear DNA was introduced into bacteria expressing the λ-Red system by electroporation, and 362 recombinant bacteria were selected on the appropriate antibiotic -containing media. When 363 necessary, the resistance marker was removed using FRT recombination mediated by the 364 expression of the Flp recombinase, as described in the same study(37). All generated mutants 365 were sequenced (Illumina, NextSeq) to confirm the deletion. The reference genome of WT 366 NILS69 can be accessed on NCBI (accession: DABQWS000000000.1) 367 368 Subcloning of defence systems and expression in E. coli K -12 MG1655 369 370 The low-copy number plasmid pZS_Cat was used to subclone NILS69 defence systems and 371 expressed them in the lab strain K -12 MG1655. Defence systems were amplified using the 372 PCR primer listed in Supplementary Table 3 , and were cloned in the pZS_Cat vector using 373 commonly used restriction digestion and ligation. Constructions were verified by Sanger 374 sequencing. The primers were designed in order to include a minimum of 200 nucleotide 375 upstream the start codon, in order to include any native promotor and regulatory sequences. 376 Only 5 out 8 defence systems could be cloned using this method. 377 378 Plasmid Vertical Transmission 379 380 The low -copy-number plasmid pZS_sfGFP was used to assess the plasmid vertical 381 transmission ability of NILS69 and its derivatives. This vector, derived from the pZ vector series 382 described by Lutz and Bujard(38) includes a low-copy-number replication origin, a kanamycin 383 resistance gene, and allows for the constitutive expression of superfolder-GFP (sfGFP) under 384 the control of an unregulated tetpromoter. 385 Vertical transmission of pZS_sfGFP was monitored as follows: overnight cultures were diluted 386 to approximately 1–10 bacteria in media containing kanamycin and incubated for 24 hours 387 (T1). The cultures were then diluted again to approximately 1 –10 bacteria in media without 388 antibiotic selection and incubated for an additional 24 hours (T2). At T1 and T2, the colony -389 forming units (CFU) were estimated by plating the cultures on solid media without selection. 390 To estimate the proportion of bacteria retaining the plasmid, ≈100 CFUs were replicated onto 391 plates containing kanamycin. 392 393 Plasmid Horizontal Acquisition 394 395 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 11 The IncC family conjugative plasmid pRCS30 was used to assess the ability of NILS69 and its 396 derivatives to acquire horizontally transferred DNA. This 15 7 kb plasmid, isolated from a 397 clinical E. coli strain, carries various antibiotic resistance genes, including aph(6) -Id, which 398 confers resistance to streptomycin. Since the spontaneous mutation frequency leading to 399 streptomycin resistance in NILS69 is below 10 ⁻⁹, resistance to streptomycin was used as a 400 marker to track the transfer of the pRCS30 plasmid. To evaluate plasmid acquisition, recipient 401 NILS69 and its derivatives, grown to an OD ₆₀₀ of 0.2, were mixed at a 1:1 ratio with the 402 pRCS30 donor strain and incubated statically. All recipient cells carried the stable high-copy-403 number plasmid pZS_sfGFP -kanR, which constitutively expressed GFP and conferred 404 resistance to kanamycin. After 1 hour and 24 hours, the mixed cultures were vortexed 405 vigorously for 1 minute, and bacteria were plated on LB agar containing kanamycin to 406 estimate the total count of recipient cells (kanR). Additionally, the cultures were plated on LB 407 agar containing both streptomycin and kanamycin to determine the number of recipient cells 408 that acquired the pRCS30 plasmid (kanR strepR). Conjugation efficiency was calculated as the 409 ratio of kanR strepR / kanR. 410 Conjugation assays across NILS derivatives were performed by first selecting spontaneous 411 mutants resistant to rifampicin (rifR), by plating the relevant strain on LB agar containing 100 412 µg/ml of rifampicin. These mutants were subsequently used as recipients during conjugation 413 assays with NILS69 derivatives carrying the pRCS30 plasmid. Here, the conjugation efficiency 414 was calculated as the ratio of rifR strepR / rifR. 415 416 Bioinformatic and data analysis tools 417 418 Defence systems from NILS69 genome were predicted using webservers of DefenseFinder 419 (6,39,40) (https://defensefinder.mdmlab.fr/) and PADLOC (22,31) 420 (https://padloc.otago.ac.nz/padloc/). Toxin Antitoxin systems were predicted with TADB 3.0 421 (https://bioinfo-mml.sjtu.edu.cn/TADB3/) 422 Search for sequence and structural homologies of the HNH endonucleases were done using 423 BLAST (41) (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE=Proteins) against the Rebase 424 database (23) (https://rebase.neb.com/rebase/rebase.html), AlphaFold3 (28) 425 (https://alphafold.ebi.ac.uk/) and FoldSeek (29) (https://search.foldseek.com/search). 426 Graphs and data analyses were generated with Prism (GraphPad). 427 428 Data accessibility 429 All raw data generated during these study have been deposited on 430 https://github.com/MMSB-MEEP-lab 431 432 Acknowledgments 433 We acknowledge Dr Luce Landraud for proving the plasmid pRCS30 and Prof. Erick Denamur 434 for sharing the strain NILS69 . The Antonina Guelin phage collection and ass ociated 435 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 12 production strains were a kind gift from Dr. Laurent Debarbieux and Dr. Aude Bernheim. AG 436 is supported by the following fundings ANR -21-CE35-0003, Emergence en recherche 2020 437 de l’Idex Université Paris Cité RM99J20IDXA8 and Emergence ville de Paris 2020 -DAE78-438 EMERGENCE. AC is supported by the ATIP -Avenir program and IdEx Université Paris Cité 439 ANR 18 IDEX 0001. 440 441 442 443

444 1. Bernheim A, Millman A, Ofir G, Meitav G, Avraham C, Shomar H, et al. Prokaryotic 445 viperins produce diverse antiviral molecules. Nature. 16 sept 2020;1-5. 446 2. Kronheim S, Daniel-Ivad M, Duan Z, Hwang S, Wong AI, Mantel I, et al. A chemical 447 defence against phage infection. Nature. déc 2018;564(7735):283-6. 448 3. Kever L, Hardy A, Luthe T, Hünnefeld M, Gätgens C, Milke L, et al. Aminoglycoside 449 Antibiotics Inhibit Phage Infection by Blocking an Early Step of the Infection Cycle. 450 mBio. 28 juin 2022;13(3):e0078322. 451 4. Rousset F, Sorek R. The evolutionary success of regulated cell death in bacterial immunity. 452 Curr Opin Microbiol. août 2023;74:102312. 453 5. Georjon H, Bernheim A. The highly diverse antiphage defence systems of bacteria. Nat 454 Rev Microbiol. oct 2023;21(10):686-700. 455 6. Tesson F, Hervé A, Mordret E, Touchon M, d’Humières C, Cury J, et al. Systematic and 456 quantitative view of the antiviral arsenal of prokaryotes. Nat Commun. 10 mai 457 2022;13(1):2561. 458 7. Maestri A, Pons BJ, Pursey E, Chong CE, Gandon S, Custodio R, et al. The bacterial 459 defense system MADS interacts with CRISPR-Cas to limit phage infection and escape. 460 Cell Host Microbe. 14 août 2024;32(8):1412-1426.e11. 461 8. Wu Y, Garushyants SK, van den Hurk A, Aparicio-Maldonado C, Kushwaha SK, King CM, 462 et al. Bacterial defense systems exhibit synergistic anti-phage activity. Cell Host Microbe. 463 10 avr 2024;32(4):557-572.e6. 464 9. Williams MC, Reker AE, Margolis SR, Liao J, Wiedmann M, Rojas ER, et al. Restriction 465 endonuclease cleavage of phage DNA enables resuscitation from Cas13-induced 466 bacterial dormancy. Nat Microbiol. mars 2023;8(3):400-9. 467 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 13 10. Birkholz N, Jackson SA, Fagerlund RD, Fineran PC. A mobile restriction-modification 468 system provides phage defence and resolves an epigenetic conflict with an antagonistic 469 endonuclease. Nucleic Acids Res. 8 avr 2022;50(6):3348-61. 470 11. Picton DM, Luyten YA, Morgan RD, Nelson A, Smith DL, Dryden DTF, et al. The phage 471 defence island of a multidrug resistant plasmid uses both BREX and type IV restriction 472 for complementary protection from viruses. 2021;18. 473 12. Dupuis MÈ, Villion M, Magadán AH, Moineau S. CRISPR-Cas and restriction-modification 474 systems are compatible and increase phage resistance. Nat Commun. 2013;4:2087. 475 13. Costa AR, van den Berg DF, Esser JQ, Muralidharan A, van den Bossche H, Bonilla BE, 476 et al. Accumulation of defense systems in phage-resistant strains of Pseudomonas 477 aeruginosa. Sci Adv. 23 févr 2024;10(8):eadj0341. 478 14. Maguin P , Varble A, Modell JW, Marraffini LA. Cleavage of viral DNA by restriction 479 endonucleases stimulates the type II CRISPR-Cas immune response. Mol Cell. 3 mars 480 2022;82(5):907-919.e7. 481 15. Srikant S, Guegler CK, Laub MT. The evolution of a counter-defense mechanism in a 482 virus constrains its host range. eLife. 4 août 2022;11:e79549. 483 16. Doron S, Melamed S, Ofir G, Leavitt A, Lopatina A, Keren M, et al. Systematic discovery 484 of antiphage defense systems in the microbial pangenome. Science [Internet]. 2 mars 485 2018 [cité 8 déc 2020];359(6379). Disponible sur: 486 http://science.sciencemag.org/content/359/6379/eaar4120 487 17. Gao L, Altae-Tran H, Böhning F, Makarova KS, Segel M, Schmid-Burgk JL, et al. Diverse 488 enzymatic activities mediate antiviral immunity in prokaryotes. Science. 28 août 489 2020;369(6507):1077-84. 490 18. Vassallo CN, Doering CR, Littlehale ML, Teodoro GIC, Laub MT. A functional selection 491 reveals previously undetected anti-phage defence systems in the E. coli pangenome. 492 Nat Microbiol. oct 2022;7(10):1568-79. 493 19. Gaborieau B, Vaysset H, Tesson F, Charachon I, Dib N, Bernier J, et al. Prediction of 494 strain level phage-host interactions across the Escherichia genus using only genomic 495 information. Nat Microbiol. nov 2024;9(11):2847-61. 496 20. Piel D, Bruto M, Labreuche Y, Blanquart F, Goudenège D, Barcia-Cruz R, et al. Phage-497 host coevolution in natural populations. Nat Microbiol. juill 2022;7(7):1075-86. 498 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 14 21. Bleibtreu A, Clermont O, Darlu P , Glodt J, Branger C, Picard B, et al. The rpoS Gene Is 499 Predominantly Inactivated during Laboratory Storage and Undergoes Source-Sink 500 Evolution in Escherichia coli Species. J Bacteriol. 15 déc 2014;196(24):4276-84. 501 22. Payne LJ, Todeschini TC, Wu Y, Perry BJ, Ronson CW, Fineran PC, et al. Identification 502 and classification of antiviral defence systems in bacteria and archaea with PADLOC 503 reveals new system types. Nucleic Acids Res. 8 nov 2021;49(19):10868-78. 504 23. Roberts RJ, Vincze T, Posfai J, Macelis D. REBASE--a database for DNA restriction and 505 modification: enzymes, genes and genomes. Nucleic Acids Res. janv 2010;38(Database 506 issue):D234-236. 507 24. Guan J, Chen Y, Goh YX, Wang M, Tai C, Deng Z, et al. TADB 3.0: an updated database 508 of bacterial toxin–antitoxin loci and associated mobile genetic elements. Nucleic Acids 509 Res. 5 janv 2024;52(D1):D784-90. 510 25. Rousset F, Depardieu F, Miele S, Dowding J, Laval AL, Lieberman E, et al. Phages and 511 their satellites encode hotspots of antiviral systems. Cell Host Microbe. 11 mai 512 2022;30(5):740-753.e5. 513 26. Hochhauser D, Millman A, Sorek R. The defense island repertoire of the Escherichia coli 514 pan-genome. PLoS Genet. avr 2023;19(4):e1010694. 515 27. Patel PH, Maxwell KL. Prophages provide a rich source of antiphage defense systems. 516 Curr Opin Microbiol. juin 2023;73:102321. 517 28. Abramson J, Adler J, Dunger J, Evans R, Green T, Pritzel A, et al. Accurate structure 518 prediction of biomolecular interactions with AlphaFold 3. Nature. 13 juin 519 2024;630(8016):493-500. 520 29. van Kempen M, Kim SS, Tumescheit C, Mirdita M, Lee J, Gilchrist CLM, et al. Fast and 521 accurate protein structure search with Foldseek. Nat Biotechnol. févr 2024;42(2):243-6. 522 30. Branger C, Ledda A, Billard-Pomares T, Doublet B, Fouteau S, Barbe V, et al. Extended-523 spectrum β-lactamase-encoding genes are spreading on a wide range of Escherichia 524 coli plasmids existing prior to the use of third-generation cephalosporins. Microb 525 Genomics. sept 2018;4(9):e000203. 526 31. Payne LJ, Meaden S, Mestre MR, Palmer C, Toro N, Fineran PC, et al. PADLOC: a web 527 server for the identification of antiviral defence systems in microbial genomes. Nucleic 528 Acids Res. 5 juill 2022;50(W1):W541-50. 529 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 15 32. Gao LA, Wilkinson ME, Strecker J, Makarova KS, Macrae RK, Koonin EV, et al. 530 Prokaryotic innate immunity through pattern recognition of conserved viral proteins. 531 Science. 12 août 2022;377(6607):eabm4096. 532 33. Picton DM, Harling-Lee JD, Duffner SJ, Went SC, Morgan RD, Hinton JCD, et al. A 533 widespread family of WYL-domain transcriptional regulators co-localizes with diverse 534 phage defence systems and islands. Nucleic Acids Res. 20 mai 2022;50(9):5191-207. 535 34. Mayo-Muñoz D, Pinilla-Redondo R, Camara-Wilpert S, Birkholz N, Fineran PC. Inhibitors 536 of bacterial immune systems: discovery, mechanisms and applications. Nat Rev Genet. 537 avr 2024;25(4):237-54. 538 35. Tesson F, Huiting E, Wei L, Ren J, Johnson M, Planel R, et al. Exploring the diversity of 539 anti-defense systems across prokaryotes, phages and mobile genetic elements. Nucleic 540 Acids Res. 9 déc 2024;gkae1171. 541 36. Kropinski AM. Measurement of the rate of attachment of bacteriophage to cells. 542

Mol Biol Clifton NJ. 2009;501:151-5. 543 37. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichia 544 coli K-12 using PCR products. Proc Natl Acad Sci U S A. 6 juin 2000;97(12):6640-5. 545 38. Lutz R. Independent and tight regulation of transcriptional units in Escherichia coli via 546 the LacR/O, the TetR/O and AraC/I1-I2 regulatory elements. Nucleic Acids Res. 15 mars 547 1997;25(6):1203-10. 548 39. Tesson F, Planel R, Egorov AA, Georjon H, Vaysset H, Brancotte B, et al. A 549 Comprehensive Resource for Exploring Antiphage Defense: DefenseFinder 550 Webservice,Wiki and Databases. Peer Community J. 25 sept 2024;4:e91. 551 40. Néron B, Denise R, Coluzzi C, Touchon M, Rocha EPC, Abby SS. MacSyFinder v2: 552 Improved modelling and search engine to identify molecular systems in genomes. Peer 553 Community J. 24 mars 2023;3:e28. 554 41. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J 555 Mol Biol. 5 oct 1990;215(3):403-10. 556 557 558 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 16 Figure and table captions 559 560 561 Figure 1. Activity of defence systems encoded in NILS69 against ten phages 562 A,B. Phage infectivity on NILS69 measured through Efficiency of Plating (EOP) assays , 563 indicated as the ratio of plaque forming unit (PFU) on NILS69 on PFU on phage production 564 strain (A) or liquid infection assays using a multiplicity of infection (MOI) of 1 (B). 565 536_P1536_P6536_P7536_P9AN24_P4DIJ07_P1LF73_P1LF73_P3LF73_P4NIC06_P2 10-8 10-6 10-4 10-2 100 102 EOP 536_P1536_P6536_P7AN24_P4DIJ07_P1LF73_P4536_P9LF73_P1LF73_P3NIC06_P2 10-1 100 101 102 103 104 105 106 EOP ΔRM (full) ΔRM_left ΔRM_right 0 1 2 3 4 5 6 7 10-4 10-3 10-2 10-1 100 Time (h) Growth OD600nm NILS69 536_P1 536_P7 536_P9 536_P6 ∆PsyrTA ∆SoFIC ∆Thoeris ∆PDC-SO4 ∆PDC-M63AB ∆PD-T7-1 ∆RM ∆PD-T4-3 ∆PD-T4-3 ∆RM ∆PD-T4-3 ∆PD-T7-1 ∆RM ∆SoFIC ∆RM ∆PD-T7-1 ∆RM ∆PD-T7-1 ∆Thoeris ∆RM ∆PD-T7-1 ∆PsyrTA ∆RM ∆PD-T7-1 ∆SoFIC ∆RM ∆PD-T7-1 ∆PDC-M63AB 536_P1 536_P6 536_P7 536_P9 AN24_P4 DIJ07_P1 LF73_P1 LF73_P3 LF73_P4 NIC06_P2 1 3 5 log10 EOP PD-T4-3PD-T7-1PsyrTAPDC-SO4 PDC-M63AB 10-6 10-4 10-2 100 102 104 EOP LM40_P1 LM40_P2 LM40_P3 CLB_P2 LF73_P1 NIC06_P2 T4LD 0 1 2 3 4 5 6 7 Time (h) NILS69 AN24_P4 DIJ07_P1 LF73_P4 0 1 2 3 4 5 6 7 Time (h) NILS69 LF73_P1 LF73_P3 NIC06_P2 MTase_I MTase_II HNH PDC-S07 SoFic PD-T4-3 PDC-M63AB PDC-S31 PD-T7-1 PDC-S07 PDC-S12 MTase_II,Vsr Thoeris_I PDC-S04 PsyrTA SoFic_2 PDC-S05 NILS69 4,808,473 bp MTase_IMTase_IIHNH endonuclease 2 HNH endonuclease 1 HNH endonuclease 3 △RM_left △RM_right WT A B C D E F .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 17 C. Genomic map of NILS69 indicating predicted defence systems (red), putative defence 566 systems assessed in this study (pink) and putative defence systems not tested in this study 567 (black). Location of prophages is indicated by yellow boxes. 568 D. Phage infectivity on NIL69 defence mutants measured through EOP assays. 569 E. Schematic of the RM island locus in NILS69 and derivative mutants. Red arrows indicate 570 genes that were flagged as defence genes by DefenseFinder and PADLOC, yellow arrows 571 indicate genes that were identified as putative endonuclease by manual inspection and black 572 arrows indicate genes encoding hypothetical proteins (upper panel). Phage infectivity was 573 measured through EOP assays (PFU on mutant NILS69/PFU on NILS69) on indicated RM 574 mutants (lower panel). 575 F. Phage infectivity on K-12 expressing NILS69 defence systems in trans from pZS plasmids. 576 The EOP was calculated as (PFU on K-12 pZS_defence plasmid)/(PFU on K-12 pZS_empty 577 plasmid). Pictures of phage T7 PFU obtained on MG1655 pZS_empty or K-12 _pZS_PD-T7-1. 578 All experiments were performed in triplicate, except for the EOP in panel A which were 579 replicated 18 times. Boxes in panel A indicate minimal to maximal values, the line indicates 580 the mean value and individual data points are shown. Error bars in panel E indicate standard 581 error of mean. 582 583 .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 18 584 585 Figure 2. NILS69 RM system limits plasmid horizontal transmission 586 A,B. Conjugation efficiency of the 157kb-plasmid pRCS30 into NILS69 and defence mutant 587 derivatives. The plasmid was transferred from its native host ( E. coli 513). Conjugation 588 efficiency is indicated as the ratio of transconjugant Colony Forming Units (CFU) on recipient 589 CFU after 1h or 24h of contact between the donor and the recipient strains. 590 C. Conjugation efficiency of pRCS30 from NILS69 to NILS69. 591 D. Vertical transmission of pZS_GFP plasmid in NILS69 and the ∆RM and ∆PD-T4-3 mutants. 592 All experiments were perfomed in triplicates, except in Panel A, conjugation in NILS69 and 593 ∆RM were performed six times. Individual data points are indicated, lines indicate mean. 594 595 596 NILS69∆PD-T4-3∆PD-T7-1∆RM ∆psyrTA ∆PDC-SO4 ∆PDC-M63AB ∆Thoeris 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 ratio transconjugant/recipient 1h 24h NILS69 x NIL69 ∆RM_left x NILS69 ∆RM_right x NILS69 NILS69 x ∆RM ∆RM_left x ∆RM ∆RM_right x ∆RM 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 Donor x Recipient ratio transconjugant/recipient 1h 24h ∆RM ∆RM_left ∆RM_right 10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 ratio transconjugant/recipient 1h 24h T0 T1 0.1 1 10 100% of plasmid loss NILS69 ΔPD-T4-3 ΔRM A B C D Donor MTase+ Donor MTase- .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 19 597 598 Supplementary material 599 600 601 Supplementary Figure 1 . Screening of the Antonina Guelin phage collection on NILS69 602 and K-12 in solid (A) and liquid (B) infection assays . 603 “nt” stands for “not tested”, the corresponding phages could not be amplified on their 604 production strain. 605 606 4h 5h 0 20 40 60 80 100 120 OD600nm variation with phage (%) AN24_P4 536_P6 536_P7 536_P9 NIC06_P2 536_P1 DIJ07_P1 LF73_P1 A B .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 20 607 0 10 20 30 102 103 104 105 106 PDP21_P1 0 10 20 30 102 103 104 105 106 PDP21_P6 0 10 20 30 103 104 105 106 107 PDP110_P2 0 10 20 30 103 104 105 106 107 PDP21_P3 0 10 20 30 103 104 105 106 107 PDP21_P8 0 10 20 30 102 103 104 105 106 T145_P1 0 10 20 30 102 103 104 105 106 PDP21_P5 0 10 20 30 102 103 104 105 106 107 PDP110_P1 0 10 20 30 102 103 104 105 106 107 T145_P2 0 10 20 30 102 103 104 105 106 T145_P3 0 10 20 30 103 104 105 106 T145_P4 0 10 20 30 104 105 106 107 T145_P5 0 10 20 30 104 105 106 107 T205_P4 0 10 20 30 103 104 105 106 107 T205_P5 0 10 20 30 103 104 105 106 107 T147_P1 0 10 20 30 103 104 105 106 107 T147_P10 0 10 20 30 102 103 104 105 106 107 T147_P5 0 10 20 30 103 104 105 106 107 T147_P7 0 10 20 30 103 104 105 106 T147_P9 0 10 20 30 103 104 105 106 107 PDP351_P2 0 10 20 30 104 105 106 107 PDP351_P3 0 10 20 30 104 105 106 107 PDP351_P4 0 10 20 30 103 104 105 106 107 536_P1 0 10 20 30 103 104 105 106 107 536_P11 0 10 20 30 104 105 106 107 536_P12 0 10 20 30 103 104 105 106 107 536_P6 0 10 20 30 103 104 105 106 107 536_P7 0 10 20 30 103 104 105 106 107 536_P9 0 10 20 30 102 103 104 105 106 107 CLB_P1 0 10 20 30 101 102 103 104 105 106 107 CLB_P2 0 10 20 30 103 104 105 106 AL505_P2 0 10 20 30 103 104 105 106 107 AL505_P3 0 10 20 30 103 104 105 106 AN24_P2 0 10 20 30 103 104 105 106 AN24_P3 0 10 20 30 103 104 105 106 107 AN24_P4 0 10 20 30 104 105 106 107 108 BCH953_P1 0 10 20 30 103 104 105 106 107 BCH953_P2 0 10 20 30 103 104 105 106 BCH953_P3 0 10 20 30 103 104 105 106 107 BCH953_P4 0 10 20 30 104 105 106 107 BCH953_P5 0 10 20 30 103 104 105 106 107 BDX03_P1 0 10 20 30 104 105 106 107 BDX03_P2 0 10 20 30 102 103 104 105 106 BDX09_P2 0 10 20 30 103 104 105 106 107 DIJ06_P1 0 10 20 30 102 103 104 105 106 107 DIJ07_P1 0 10 20 30 102 103 104 105 106 LF110_P1 0 10 20 30 102 103 104 105 106 107 108 LF110_P2 0 10 20 30 103 104 105 106 107 LF110_P3 0 10 20 30 104 105 106 107 LF110_P4 0 10 20 30 102 103 104 105 106 107 LF31_P1 0 10 20 30 101 102 103 104 105 106 LF31_P3 0 10 20 30 102 103 104 105 106 LF50_P3 0 10 20 30 103 104 105 106 107 LF7074_P1 0 10 20 30 101 102 103 104 105 106 LF7074_P2 0 10 20 30 103 104 105 106 107 LF7074_P3 0 10 20 30 104 105 106 107 LF73_P1 Free phage (PFU/ml) Time (min) .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 21 608 609 Supplementary Figure 2. Adorption assays of 93 phages from the Antonina Guelin on 610 NILS69 and on their control strains (phage production strains) 611 612 Free phage (PFU/ml) Time (min) 0 10 20 30 103 104 105 106 107 108 LF73_P3 0 10 20 30 103 104 105 106 107 LF73_P4 0 10 20 30 103 104 105 106 107 LF82_P1 0 10 20 30 102 103 104 105 106 LF82_P2 0 10 20 30 103 104 105 106 LF82_P3 0 10 20 30 103 104 105 106 LF82_P4 0 10 20 30 102 103 104 105 106 LF82_P5 0 10 20 30 103 104 105 106 LF82_P6 0 10 20 30 103 104 105 106 107 LF82_P8 0 10 20 30 101 102 103 104 105 106 107 LF82_P9 0 10 20 30 103 104 105 106 LI10_P1 0 10 20 30 102 103 104 105 106 107 LI10_P2 0 10 20 30 103 104 105 106 107 LI10_P3 0 10 20 30 103 104 105 106 107 LI10_P4 0 10 20 30 103 104 105 106 107 LI10_P5 0 10 20 30 103 104 105 106 LI10_P6 0 10 20 30 101 102 103 104 105 106 LM02_P1 0 10 20 30 101 102 103 104 105 106 LM07_P1 0 10 20 30 103 104 105 106 LM08_P1 0 10 20 30 104 105 106 LM08_P2 0 10 20 30 103 104 105 106 107 LM33_P1 0 10 20 30 102 103 104 105 106 LM40_P1 0 10 20 30 102 103 104 105 106 107 LM40_P2 0 10 20 30 102 103 104 105 106 107 LM40_P3 0 10 20 30 103 104 105 106 MT1B1_3A1 0 10 20 30 102 103 104 105 106 107 NAN33_P1 0 10 20 30 102 103 104 105 106 107 NAN33_P2 0 10 20 30 101 102 103 104 105 106 107 NAN33_P4 0 10 20 30 102 103 104 105 106 107 NAN33_P5 0 10 20 30 102 103 104 105 106 NAN33_P6 0 10 20 30 101 102 103 104 105 NIC06_P2 0 10 20 30 104 105 106 107 NIC06_P3 0 10 20 30 104 105 106 NRG857C_P1 0 10 20 30 103 104 105 NRG857C_P2 0 10 20 30 102 103 104 105 106 107 NRG857C_P3 0 10 20 30 103 104 105 106 107 T4LD 0 10 20 30 103 104 105 106 T7LD .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 22 613 614 615 Supplementary Figure 3. Growth kinetic of NILS69 and mutant dertivative in the presence 616 of the infective phages at a MOI of 1 617 618 1 2 3 4 5 6 7 10-2 10-1 100 536_P1 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 AN24_P4 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 536_P6 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 536_P7 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 536_P9 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 DIJ07_P1 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 LF73_P4 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 LF73_P1 Time (h) Growth OD600nm 1 2 3 4 5 6 7 10-2 10-1 100 Time (h) Growth OD600nm NIC06_P2 1 2 3 4 5 6 7 10-2 10-1 100 Time (h) Growth OD600nm LF73_P3 ΔPsyrTA ΔThoeris NILS69 WT ΔPD-T4-3 ΔPD-T7-1 ΔRM ΔSoFic .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 23 619 620 Supplementary Figure 4: Predicted structure of HNH endonuclease 3 621 A. HNH endonuclease 3 encoded in NILS69 has 100% identity with the amino acid sequence 622 of a HNH endonuclease (Uniprot A0A3K0QCZ9), which has a predicted structure in AlphaFold 623 (average pLDDT score of 80,56) 624 B. Structural alignment of NILS69_HNH endonuclease and K-12 MG1655_EcoKMrcA (e-value 625 2.48e-4; sequence identity 20.6%) 626 627 628 A B .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint 24 629 Supplementary Figure 5: Impact of bacterial lawn density on LM40_P phage infectivity in 630 the K12 MG1655 strain expressing PsyrTA and its parental strain. 631 A, B. The bacterial lawn was inoculated directly from cultures at OD600 = 0.6 (1x). 632 C, D. The parental and PsyrTA strains were inoculated at dilutions of 1/10 or 10x, respectively, 633 from cultures at OD600 = 0.6. For each infectivity test, 3 µL of 10 -fold serially diluted phage 634 was spotted. The T4 phage was used as a positive control for lysis. 635 636 637 Supplementary Table 1. List of phages infecting NILS69 638 639 Supplementary Table 2. List of predicted defence systems and toxin antitoxin systems 640 encoded in NILS69 641 642 Supplementary Table 3. Lists of plasmids, strains and primers used in this study 643 pZS_PsyrTA 10X pZS_empty 1X1X P1 P2 P3 T4 LM40 P1 P2 P3 T4 LM40 0,1X A D B C .CC-BY-NC-ND 4.0 International licenseperpetuity. It is made available under a preprint (which was not certified by peer review) is the author/funder, who has granted bioRxiv a license to display the preprint in The copyright holder for thisthis version posted December 21, 2024. ; https://doi.org/10.1101/2024.12.20.629700doi: bioRxiv preprint