Prophages block cell surface receptors to ensure survival of their viral progeny

Abstract

In microbial communities, viruses compete for host cells to infect, and thus evolve diverse ways to inhibit their competitors. One mechanism is Superinfection exclusion (Sie), whereby a virus that has established an infection prevents a secondary infection. This phenomenon has been shown to play an important role in the spread of eukaryotic viruses. In this work, we determine that bacterial virus (phage) Sie proteins perform a similar role, promoting viral spread through the bacterial community. We characterize a Sie system that inhibits superinfection by altering the dynamics of a common cell surface receptor, the type IV pilus. This protein, known as Zip, does not abrogate pilus activity, but fine tunes it, providing strong phage defence without a fitness cost. Notably, Zip activity prevents internalization and destruction of phage progeny, which we call the “anti-Kronos effect” after the Greek god who devoured his children. We show that anti-Kronos activity greatly increases the number of free phages in a lysogen community, and that it is a conserved feature of diverse prophage-encoded Sie systems. Our results reveal the mechanistic basis of a novel Sie system, show that it functions to ensure survival of phage progeny, and provide new insight into the conservation of viral defence mechanisms amongst bacterial and eukaryotic systems.

The ability of an established viral infection to inhibit subsequent infection by a closely related virus can have profound effects on viral evolution and the structure of viral populations¹. This phenomenon, known as Superinfection exclusion (Sie), was first characterized in phages, the viruses that infect bacteria². It has subsequently been described for a wide range of viruses that infect animals, plants, and bacteria^3–8. While genes that endow Sie are a common feature of many viruses, and detailed mechanisms of activity have been determined for many systems, their evolutionary significance has remained largely unclear. Several theories have been put forth – one possible advantage is to protect the infected cell from other viruses that would compete for host cell resources. Alternatively, it has been proposed that Sie may permit cooperation between the same or closely related viruses, with the exclusion of related viruses benefiting the community by allowing them to efficiently search out uninfected cells⁹.

Phages have been shown to mediate superinfection exclusion via mechanisms that inhibit the adsorption of competing phages to the host cell surface or prevent the successful injection of their genomes across the bacterial cell envelope. For example, in E. coli, phage T5 produces a lipoprotein early in infection that prevents further adsorption events by blocking the cell surface receptor protein, FhuA¹⁰. Phage T4 encodes proteins that block superinfection by T-even phages by inhibiting phage lysozyme activity¹¹, and HK97 expresses a small protein that inhibits the DNA injection process of superinfecting phages^12,13. P. aeruginosa phages have been shown to block cell surface receptors by modifying the O-antigen^14,15 and interfering with pilus assembly^16,17. As many P. aeruginosa phages rely on the type IV pilus for infection, pilus inhibition provides a useful mechanism for blocking further phage infection. However, this is a double-edged sword; while the disruption of pilus assembly protects against superinfection, P. aeruginosa that lack type IV pili are unable to adhere to surfaces and establish beneficial biofilms or escape unfavourable environments^18–20. How prophages that inhibit pilus assembly to provide superinfection exclusion balance the fitness advantage of phage resistance with the potential cost associated with the loss of the pilus is unknown.

In a previous study of P. aeruginosa phages, we identified a protein encoded by phage JBD26 that abrogated twitching motility and endowed strong superinfection exclusion activity against all phages that require the type IV pilus for infection²¹. This protein, which we call Zip (PilZ interacting protein; YP_010299255), was shown to be highly expressed from the prophage during late exponential growth²². However, in contrast to the complete abrogation of twitching motility observed when Zip was expressed from a plasmid, expression from its natural context in the JBD26 prophage decreased twitching motility only ~20% as compared to wild-type PA14, while still providing very strong phage resistance^21,22. This paradox of high resistance with minimal effects on twitching motility motivated us to investigate this system in detail to determine how the prophage manipulates the bacterial cell to increase fitness via phage resistance while avoiding the evolutionary cost of loss of the pilus. This work also allowed us to address the question of why providing superinfection exclusion at the bacterial cell surface is so advantageous for temperate phages.

Zip binds to PilZ

To determine the mechanism by which Zip provides phage defence, we first sought to identify its bacterial target. As Zip expression inhibited twitching and mediated strong resistance to phages that require the type IV pilus for infection, we performed a bacterial two-hybrid assay²³ with a library expressing components of the P. aeruginosa type IV pilus. This library included major and minor pilins, chaperones, scaffold, and motor protein genes. We detected a strong interaction between Zip and PilZ (PA2960; Extended Data Fig. 1a). Previously, a pilZ deletion was shown to prevent type IV pilus assembly, thereby abrogating twitching motility and increasing phage resistance²⁴. These phenotypes are consistent with those observed with Zip expression, suggesting that PilZ is a biologically relevant target.

P. aeruginosa type IV pili can be classified into five different families based on the sequence of the major pilin subunit, PilA, and the presence or absence of associated major and minor pilin proteins²⁵. PilZ shares 100% sequence identity across these five families, suggesting that Zip activity should inhibit all of them. To test this, we expressed Zip from a plasmid in five strains that each encode a different pilus type and determined that it abrogated twitching motility across all five families (Fig. 1a). We next overexpressed PilZ from a plasmid in cells containing a JBD26 prophage to see if this could rescue the ability of phages to plate by ensuring that excess PilZ was present in the cell to overcome the Zip-mediated inhibition. We found that overexpression of PilZ in the lysogen increased phage susceptibility by >10⁴-fold (Fig. 1b), providing further evidence that the highly conserved PilZ is the target of Zip activity.

Fig. 1 |. Zip targets a conserved pilus chaperone, PilZ.

a, Twitching motility of cells expressing Zip from a plasmid. b, Ten-fold serial dilutions of phage JBD30 plated on lawns of PAO1, the JBD26 lysogen, JBD26 lysogen expressing PilZ from a plasmid, and PAO1 expressing Zip-GFP. A representative image of three replicates is shown. c, SDS-PAGE gel analysis of co-purifications of Zip, PilZ and PilB stained with Coomassie blue. d, Pilus assembly and retraction results from the activity of the PilB assembly ATPase, which incorporates PilA monomers into the growing filament, and the PilT ATPase, which disassembles the filament and causes pilus retraction. e. Fluorescence microscopy of cells expressing GFP-tagged Zip (top), or GFP alone (bottom). f, Total number of pili individual cells make in a 30 second time interval. g, Percent of cells that produce at least one pilus in 30 secs after overnight growth. Significance of a one-way ANOVA (** p<0.005) shown. h – j Pilus dynamics of mid-log cells (OD = 0.3). Data are representative of three biological replicates. h, Maximum length of individual pili. Significance of a two-tailed Mann-Whitney test (*p<0.01, ***p<0.001) i, The extension velocity of individual pili. j, Total time individual pili spend extending. Significance of a two-tailed Mann-Whitney test (***p<0.05) k, Length of pili individual cells make in a 30 second time interval at high cell density (overnight culture). Data are representative of three biological replicates. Results of a one-way ANOVA (* p<.01, **** p<0.0001).

PilZ, a known regulator of type IV pilus function, has previously been shown to bind the PilB assembly ATPase protein that is required for type IV pilus biogenesis. This interaction is thought to stabilize the PilB N-terminal domain and allow a hexamer of PilB to efficiently localize to the pilus scaffold complex^26,27. Once attached to the pilus scaffold, the PilB hexamer hydrolyses ATP and provides the energy to incorporate hundreds of PilA subunits into the extending pilus. To determine if Zip affects the ability of PilZ to chaperone PilB, we used protein co-purification experiments to examine the complexes. First, we confirmed the direct interaction between Zip and PilZ by co-expressing His-tagged Zip with untagged PilZ in E. coli and performing a Ni-affinity chromatography co-purification experiment. We found that untagged PilZ co-eluted with the His-tagged Zip, confirming direct protein binding (Fig. 1c). We next co-expressed His-tagged PilZ and untagged PilB and detected a stable complex (Fig. 1c). To determine if Zip interferes with this interaction, we co-expressed His-tagged Zip together with untagged PilZ and PilB and performed Ni-affinity purification. We found that all three proteins co-purified, suggesting the formation of a tripartite complex (Fig. 1c). To further characterize this complex and determine if Zip binds directly to PilB or if this interaction is mediated through PilZ, we co-expressed His-tagged Zip with untagged PilB and found that they did not co-purify (Fig. 1c). Together, these data imply that Zip interacts with PilZ, and that Zip does not prevent PilZ from binding to PilB.

Zip localizes to the cell poles and makes shorter pili

The PilZ-PilB interaction is known to allow efficient localization of the PilB ATPase to the pilus scaffold complex located at the poles of the cell²⁷. The activity of Zip could interfere with this localization, or it could affect the ability of the PilB motor to incorporate the PilA subunits into the extending pilus filament (Fig. 1d). To determine if Zip localizes to the poles of the cell where the pilus complex is known to assemble, we expressed green fluorescent protein (GFP)-tagged Zip with its native promoter from a plasmid in PAO1 and examined its localization using fluorescence microscopy. First, to ensure that the fusion of GFP to the C-terminus of Zip did not interfere with its activity, we assessed twitching motility and phage resistance of cells expressing the fusion protein. We found that Zip-GFP expression in PAO1 resulted in a decrease in twitching of approximately 20% as compared to wild-type PAO1, a level comparable to the twitching inhibition observed with the JBD26 lysogen (Extended Data Fig. 1b). Zip-GFP expression also provided strong phage resistance (Fig. 1b) showing that the GFP tag did not inhibit activity. Examination of these cells using fluorescence microscopy revealed that Zip-GFP accumulated at the poles of the cells while GFP alone was diffusely located throughout the cytoplasm (Fig. 1e). These data show that Zip localizes to the poles of the cells where the type IV pilus machinery is known to assemble.

While the co-purification experiments confirmed that PilZ is the target of Zip binding, it was not clear how this interaction would affect type IV pilus assembly. To gain insight into Zip activity in the context of the prophage, we used fluorescence microscopy to directly examine pili on the surface of cells that contain either a wild-type JBD26 prophage or one lacking zip (JBD26^Δzip). In these experiments, the major pilin subunit (PilA) was fluorescently labelled with thiol-reactive maleimide dye Alexa88-mal²⁸. We first compared the number of pili present on the surface of the cells at mid-log phase (OD ~0.3) and following overnight growth. We found no significant difference between wild-type PAO1, the JBD26 lysogen or the JBD26^Δzip lysogen during early exponential phase; all cells in each of the three samples were piliated, with an average of ~4 pili assembled per cell per 30 second increment (Fig. 1f). By contrast, following overnight growth we found that ~60% of JBD26 lysogens had no pili on their surface, while 100% of the wild type PAO1 and the JBD26^Δzip lysogen were piliated (Fig. 1g). Notably, there were also fewer pili on the surface of these cells than at mid-log phase growth, with more than half of the PAO1 and JBD26^Δzip lysogen cells displaying only a single pilus as compared to four at the earlier timepoint.

To gain additional insight into how Zip activity affected pilus assembly, we measured the length of pili on the surface of cells, as well as their extension and retraction velocities. Using the early exponential phase cells where all cells displayed at least one pilus, we first calculated the distribution of pilus lengths. We found that the JBD26 lysogen had a median length of 0.4 μm, while wild-type PAO1 and JBD26^Δzip lysogen showed median lengths of 0.6 μm (Fig. 1h). To determine why the pili were shorter on average in the JBD26 lysogen, we next examined extension dynamics of individual pili. While the extension velocity was similar across the three strains (Fig. 1i), the extension time was shorter for the JBD26 lysogen (Fig. 1j). The current model for pilus assembly suggests that only one type of motor, PilB extension or PilT retraction, can be bound to the pilus machinery at a time²⁹. In addition, the binding and unbinding of these motors is stochastic, with the duration of each pilus extension event proportional to how quickly the PilB extension motor becomes unbound³⁰. Thus, our data suggest that Zip activity destabilizes the interaction of PilB with the pilus machinery, causing the extension motor to disassociate from the scaffold earlier and thereby generate shorter pili. As we had evidence that Zip activity was increased following overnight growth, with only 60% of JBD26 lysogens displaying pili in the 30 second timeframe we examined, we also examined the length of the pili on these cells (Fig. 1k). We found that the JBD26 lysogen pili were shorter on average, at a length of 0.2 μm as compared to 0.4 μm for PAO1. Together, these data imply that Zip activity does not prevent pilus assembly, but instead modulates it in a way that decreases the length of pili produced. In addition, at high cell density Zip expression causes complete loss of pili in almost half the bacterial community (Fig. 1f).

Zip expression is controlled by LasR, the principal quorum sensing regulator

Our previous work showed that zip is the most highly expressed gene in the JBD26 prophage during late exponential phase²². As it appeared to be expressed independently from other prophage genes, we looked for a potential promoter and discovered a 74-nucleotide non-coding region upstream of the gene. To determine if this sequence includes a promotor that drives zip expression, we cloned it into the pQF50 reporter plasmid³¹, which placed it upstream of lacZ and allowed its activity to be monitored via β-galactosidase activity. We transformed this plasmid into wild-type PA14 and observed robust β-galactosidase activity at 18 hours (Fig. 2a). As this experiment was done in the absence of a JBD26 prophage, it suggested that expression from the JBD26 zip promoter not reliant on a phage transcriptional regulator. To determine if a bacterial transcriptional regulator was mediating expression from the zip promotor, we assayed β-galactosidase activity in PA14 strains containing single gene deletions of various transcriptional regulators. We found that a culture of the lasR mutant (ΔlasR), the principal regulator of quorum sensing, showed greatly decreased β-galactosidase activity following overnight growth (Fig. 2a). Expression in the rhlR mutant was partially attenuated (Fig. 2a). As RhlR and LasR are known to activate transcription by binding to a DNA sequence called the las-box with different affinities, finding that both regulators affected zip transcription but to different degrees was consistent with previous studies^32–34. By contrast, mutants in other quorum regulators such as vfr and pilR did not show a significant difference in β-galactosidase activity as compared to wild-type PA14, suggesting that they are not involved in activation of gene expression associated with this promoter sequence. To provide additional evidence that LasR was driving β-galactosidase expression, we transformed the reporter construct into a QteE deletion strain (ΔqteE). QteE reduces the stability of LasR, and its absence allows more LasR to accumulate at lower cell density³⁵. We observed increased β-galactosidase activity during exponential growth phase in the ΔqteE mutant as compared to wild-type PA14 (Fig. 2b), consistent with LasR activating expression from the promoter. These data imply that the zip promoter is controlled by the bacterial quorum sensing regulator, LasR. To confirm that LasR controls zip expression from the prophage, we tested the ability of the JBD26 prophage to resist infection by the pilus-dependent phage, JBD30; while the JBD26 prophage in wild-type PA14 robustly resisted JBD30 infection, it showed >1,000-fold increased susceptibility to JBD30 infection in the ΔlasR strain (Fig. 2c). These data show that P. aeruginosa prophages exploit the host cell quorum sensing pathway to drive expression of anti-phage defence.

Fig. 2 |. LasR drives Zip expression.

a, β-galactosidase activity resulting from gene expression driven by the Zip promoter. Single gene deletions in various PA14 transcriptional regulators are shown. Three independent biological replicates are shown. Statistical significance was calculated using a one-way ANOVA (****p< 0.0001). b, β-galactosidase activity of the Zip promoter measured during late-exponential growth in PA14, ΔlasR and ΔqteE strains. Three biological replicates are shown. Significance by a one-way ANOVA (**p < 0.005). c, Serial dilutions of phage JBD30 spotted on a wild type JBD26 lysogen, a JBD26^Δzip lysogen, and a wild type JBD26 lysogen in a ΔlasR background. A representative image of three biological replicates is shown. d, β-galactosidase activity driven by the zip promoter from JBD26 and JBD24 in PA14 and the ΔlasR strain. Three biological replicates are shown. Significance by a two-way ANOVA (*** p< 0.005, **** p<0.0001). e, Serial dilutions of phage JBD30 spotted on lawns of PA14, lysogens of: JBD26, JBD26^Δzip, JBD24, JBD24^Δzip and ΔlasR–JBD24, either immediately after pouring (0 hours) or following incubation to allow cell growth (4 hours). An image representative of three independent biological replicates is shown.

We had previously noted that a JBD24 lysogen, which is closely related to JBD26 and encodes a homologue of Zip that shares 95% sequence identity (Gp58; YP_007392821), was sensitive to infection by most pilus-dependent phages²¹. A sequence alignment revealed that the promoter region upstream of JBD24 gene 58 differed from the JBD26 sequence upstream of zip, sharing 71% sequence identity (Extended Data Fig. 2a). To determine if this promoter sequence was also responsive to LasR we cloned it into the pQF50 reporter plasmid upstream of lacZ and assayed its activity using a β-galactosidase assay. This promoter showed much weaker activity than the zip promoter from JBD26, but still appeared responsive to LasR (Fig. 2d). This weaker promoter activity provides an explanation for the lower superinfection exclusion activity previously noted for the JBD24 lysogen.

To determine if quorum sensing might upregulate expression of Zip and provide JBD24 with greater protection from superinfection at high cell density, we applied a phage JBD30 lysate onto a JBD24 lysogen bacterial lawn that was pre-incubated at 30 °C for several hours to allow the quorum sensing system to activate. We found that JBD30 was able to form plaques on the JBD24 lysogen when phages were applied immediately. By contrast, when the JBD24 lysogen lawn was grown for 4 hours before the phages were applied, complete inhibition of plaquing was observed (Fig. 2e), suggesting that defence was upregulated at high cell density. To determine if this effect was due to the activity of the Zip protein, we created a mutant phage lacking gene 58 (JBD24^Δzip) and repeated the experiment. We found that the ability of the JBD24^Δzip lysogen to resist infection by phage JBD30 was greatly diminished when the phage lysate was applied both immediately and after 4 hours of cell growth (Fig. 2e). Similarly, a JBD24 prophage in a ΔlasR strain background was also not able to defend against phage infection at high cell density (Fig. 2e). These data show that the JBD24 Zip protein does provide effective Sie protection, and that this activity is activated at high cell density in a LasR-dependent manner. This contrasts with JBD26, where Zip defence is active at both low and high cell densities (Fig. 2e).

The Kronos effect: lysogens destroy their viral progeny in the absence of Zip

JBD26 prophages spontaneously induce such that an overnight culture of a JBD26 lysogen grown under normal conditions contains abundant viral particles to a level of ~10⁶ plaque forming units (pfu)/mL. Remarkably, we discovered that an overnight culture of a JBD26^Δzip lysogen contained fewer than 500 pfu/mL. To investigate this phenomenon, we monitored the number of phage progeny produced by lysogens of JBD26 and JBD26^Δzip over time. To do this, we first removed any free phages from overnight cultures of JBD26 and JBD26^Δzip by collecting the cells by centrifugation and washing several times. We then grew the cultures and took samples at 3, 6, 9, 12 and 20 hours after the start of cell growth and titered the number of phages at each time point. We found that while JBD26 and JBD26^Δzip lysogen cultures accumulated similar numbers of phages in the supernatant at both 3 and 6 hours, after 20 hours the titer of JBD26^Δzip was only ~500 pfu/mL as compared to ~10⁶ pfu/mL in the wild-type JBD26 lysogen culture (Fig. 3a). Notably, between 9 and 20 hours the JBD26^Δzip lysogen titer fell from ~10⁵ pfu/mL to 500 pfu/mL, showing that phages were being actively destroyed. To confirm that the observed difference in phage numbers was due to the loss of Zip activity and not a phage induction or replication defect, we treated the two lysogens with mitomycin C, a DNA damaging agent that causes induction of most of the prophages and subsequent cell lysis. We found that both JBD26 and JBD26^Δzip prophages produced titers of 10⁶ pfu/mL (Fig. 3b), showing that phage replication was not affected by zip deletion. The difference in the number of phage progeny at 20 hours was also not due to differences in bacterial population sizes as both JBD26 and JBD26^Δzip cultures grew to 4×10⁹ colony forming units (cfu)/mL at 20 hours (Fig. 3c). Since both JBD26 and JBD26^Δzip lysogen cultures accumulated equal numbers of phages at 6 hours, we postulated that the loss of phages at the 20-hour timepoint might be due to the JBD26^Δzip lysogens not being able to block infection at the cell surface via the activity of Zip, as does the wild-type JBD26 lysogen. This could lead to reinfection by the phage progeny spontaneously produced by the bacterial lysogen. To determine if this were true, we repeated the growth experiment with the same lysogens containing a plasmid expressing Zip. We found that when Zip was supplied from a plasmid, the JBD26 and JBD26^Δzip lysogens displayed equal numbers of free phages in overnight cultures. This shows that the loss of Zip activity is the reason for the low titer observed in the JBD26^Δzip lysogen (Fig. 3d).

Fig. 3 |. The anti-Kronos effect promotes survival of viral progeny.

Phage titers a, spontaneously produced by JBD26 and JBD26^Δzip lysogens over time and b, following mitomycin C treatment. c, Number of bacterial cells present in JBD26 and JBD26^Δzip lysogen cultures over time. d, Phage production by JBD26 and JBD26^Δzip lysogens in the absence and presence of Zip supplementation from a plasmid. e, The anti-Kronos effect provides the JBD26 lysogen with superinfection exclusion activity that vastly increases the number of phage progeny that accumulate in the community. f, Phage titers spontaneously produced by JBD24 and JBD24^Δzip over time and g, Number of bacterial cells present in JBD24 and JBD24^Δzip lysogen cultures over time. h, Phage production by JBD24 and JBD24^Δzip lysogens in the absence and presence of Zip supplementation from a plasmid. For all panels, data are representative of three independent biological replicates. Statistical significance was measured using two-way ANOVA, significant p-values are noted (***p<0.005, ****p < 0.0001).

Taken together, these data indicate that Zip performs a crucial function within a lysogenic culture by preventing phage particles produced through spontaneous prophage induction from adsorbing to and injecting their DNA into other lysogenic cells (Fig. 3e). As only a small percent of cells in the community are spontaneously producing phage, the chance of the progeny adsorbing to and infecting other lysogens is high. Even though these phage infection events would not kill the lysogens due to immunity conferred by the prophage-expressed repressor protein, which represses phage transcription when the genome enters the cell, the unproductive infections lead to abundant loss of viable phage particles. This effect is shown by the low number of phage progeny observed in the JBD26^Δzip lysogen culture at 20 hours (Fig. 3a). Zip expression prevents these unproductive infections by preventing phage adsorption to the pilus, and thereby avoiding futile injection of phage DNA into lysogenic cells (Fig. 3e). This phenomenon vastly increases the number of free phages in the community that could go on to infect other non-lysogenic hosts. We have named the phenomenon of lysogenic cells destroying their own viral particles through surface adsorption and DNA injection as the “Kronos effect” after the Greek Titan Kronos, who ate his own children. Thus, prophage-encoded proteins that block phage adsorption at the cell surface can be referred to as anti-Kronos factors.

While the JBD24 Zip protein did not endow phage resistance at low cell density (Fig. 2e), we wondered if it would show a similar anti-Kronos effect. To investigate this, we assessed the number of free phages present in overnight cultures of JBD24 and JBD24^Δzip lysogens. We found that the wild-type JBD24 lysogen accumulated ~10⁵ pfu/mL, while the JBD24^Δzip lysogen produced only ~100 pfu/mL (Fig. 3f), despite both cultures growing to equal cell densities (Fig. 3g). Like JBD26^Δzip, the JBD24^Δzip lysogen showed phage production equal to the wild type lysogen at 6 hours. At 9- and 12-hours, the titer of JBD24^Δzip remained constant, while the wild-type phage continued to increase, with ~100-fold more phages/mL accumulating in the wild-type JBD24 culture at the 12hour timepoint. While titers of both phages decreased overnight, the titer of wild type JBD24 remained 1,000-fold higher than JBD24^Δzip (Fig. 3f). To determine if higher Zip expression levels would result in the accumulation of more phages in these cultures, we repeated the experiment with the lysogens containing a plasmid expressing Zip. We found that the phage titers of both lysogens increased to ~10⁶ pfu/mL (Fig. 3h), implying that the low titers observed were a result of the Kronos effect. It is interesting to note that the phage titers of JBD24 produced in the presence of plasmid-expressed Zip were equal to the phage titer produced by the wild-type JBD26 lysogen (Figures 3d,h). The lower phage titer naturally produced by the JBD24 lysogen is likely the result of the lower expression levels provided by the native zip promoter found in this phage (Fig. 2d), resulting in a stronger Kronos effect. These data confirm that the Zip homologue in JBD24 is also required for stable accumulation of free phages in the community.

The zip homologues were found in the same genomic position in the closely related phages JBD26 and JBD24. To determine how widespread homologues of this protein are, we performed a PSI-BLAST search starting with the JBD26 Zip sequence. We identified ~1000 homologues in a variety of P. aeruginosa strains and phages, with 87 of the 1038 Pseudomonas phage genomes present in the NCBI database encoding one, suggesting that this is a common function. Alignment of these protein sequences showed high conservation among the homologues, with pairwise sequence identities above 80% (Extended Data Fig. 2b). While most phage homologues were found in the same genomic position as zip, immediately following the late gene operon, some were found at the other end of the late gene operon, upstream of the small terminase gene (Extended Data Fig. 2c). Regardless of the genomic position, the phage homologues maintained the 5’ upstream region corresponding to the zip promoter, and all contained the predicted Las-box motif (Extended Data Fig. 2a). These data show that Zip is a conserved feature of P. aeruginosa phages and that it provides a widespread mechanism through which lysogens can ensure that phage progeny are protected from loss via the anti-Kronos effect.

The anti-Kronos effect is conserved and mediated through diverse mechanisms

As the anti-Kronos effect provides such a striking potential evolutionary advantage by greatly increasing the number of phage progeny that can accumulate within a lysogenic community and go on to infect new hosts, we wondered if this effect is widespread and can be mediated by diverse mechanisms. We surveyed the literature to find previously characterized Sie mechanisms that act at the cell surface and identified eight additional systems (Extended Data Table 1). These systems are found across a broad range of Gram-positive and -negative bacteria, and act through mechanisms that include type IV pilus modification, serotype conversion, techoic acid modification, and blocking access to membrane protein receptors. We selected another pilus-mediated Sie protein, found in P. aeruginosa phage LESϕ3 for analysis. This protein, encoded by gene 50 and referred to here as gp50, interacts with the type IV pilus machinery to block infection by LESϕ3 and other related phages. Gp50 is not related to Zip and inhibits phage attachment by a mechanism that is completely distinct from the Zip mechanism³⁶. Like zip, gene 50 has a predicted untranslated region at its 5’ end, suggesting that its expression is likely driven by its own promoter (Fig. 4a, Extended Data Fig. 2d). To determine if gp50 also protects phage progeny from the Kronos effect, we assessed the number of phages that accumulated in an overnight culture of a mutant LESϕ3 lysogen that lacks gene 50 (LESϕ3^Δ50). We found that the LESϕ3^Δ50 lysogen produced an equal number of phage progeny as wild-type LESϕ3 when induced with mitomycin C (Fig. 4b, right), showing that phage replication and assembly were not affected by this gene deletion. We next examined the number of phage progeny spontaneously produced by the lysogens at 3, 6, and 20 hours and found that there was a 10⁴-fold decrease in free phage progeny in the LESϕ3^Δ50 culture at 20 hours (Fig. 4b, left). These results show that the anti-Kronos effect is also manifested in this system.

Fig. 4 |. Anti-Kronos systems are widespread.

a, Phage LESϕ3 encodes a Sie protein between the tail tube and tail assembly chaperone (TAC). b, Phage titers spontaneously produced by LESϕ3 and LESϕ3^Δ50 over time (left) and following mitomycin C induction (right). c, Phage JBD44 encodes a two gene Sie system between the endolysin and tail spike proteins. d, Phage titers spontaneously produced by JBD44 and JBD44^Δ40–41 over time (left) and following mitomycin C induction (right). e, Phage ε15 from Salmonella anatum encodes a serotype conversion unit between the tail fiber and holin. f, Phage titers spontaneously produced by ε15 and ε15^ΔwzyB-iap over time (left) and following mitomycin C induction (right). g, Phage HK97 encodes a Sie, gp15, between the tail protein and the tape measure protein. h, Phage titers spontaneously produced by HK97 and HK97^Δ15 over time (left) and following mitomycin C induction (right). For all panels, data are representative of three independent biological replicates. Statistical significance was measured using two-way ANOVA, significant p-values are noted (****p < 0.0001, ***p<0.005). Genetic locus of each anti-Kronos shown here present in Extended Fig. 2d.

Not all P. aeruginosa phages use the type-IV pilus for infection. We wondered if phages that rely on the other major cell surface receptor, lipopolysaccharide (LPS), also modify their surface to protect against the Kronos effect. Phage JBD44, which is known to use LPS as a cell surface receptor²¹, encodes two genes whose activity acetylates the bacterial O-antigen and provides an LPS modification known as serotype O10 to the host in which the prophage resides³⁷. These genes are located at the end of the tail operon, between the tail spike and endolysin genes (Fig. 4c, Extended Data Fig. 2d). We deleted the genes corresponding to these LPS-modifying enzymes from JBD44 (JBD44^Δ40–41) and monitored phage production over time. We found that the JBD44^Δ40–41 lysogen culture accumulated ~10⁴-fold fewer phages/mL in the culture medium than wild-type JBD44 after overnight incubation (Fig. 4d, left). This was not due to problems with phage replication or assembly of the virion as the mutant phage produced the same number of phage progeny following mitomycin C induction (Fig. 4d, right). In contrast to JBD26 and LESϕ3, the Kronos effect is strongly observed at all timepoints. This is likely a result of the different receptors being used. Both JBD26 and LESϕ3 use the type IV pilus, which is not highly expressed in cells growing in liquid medium, and thus does not provide many receptors for phages to bind. By contrast, JBD44 uses LPS, which coats the entire cell and provides an abundance of receptor with which the phages can interact and thereby be lost through the Kronos effect. These data show that JBD44 also expresses anti-Kronos proteins that protect spontaneously produced phage progeny in the community from being lost due to reinfection of the lysogen.

To determine if the ability to inhibit self-infection at the cell surface is a general feature of P. aeruginosa prophages, we assessed the capacity of lysogens to resist infection by their own phages. We selected five diverse P. aeruginosa phages (Extended Data Fig. 3a) for which we could detect strong adsorption within five minutes and created individual lysogens of them. We then mixed each lysogen with its phage at a multiplicity of infection of 0.1 and quantitated phage adsorption. We found that all five showed a decrease in free phages in the supernatant when mixed with wild-type cells, illustrating that adsorption and infection was occurring (Extended Data Fig. 3b). By contrast, when these phages were mixed with their cognate lysogen, no decrease in phage titer was observed at five minutes (Extended Data Fig. 3b). These results suggest that blocking phage adsorption at the cell surface is a common feature of P. aeruginosa phages.

To discover if the anti-Kronos effect is conserved in prophages outside of P. aeruginosa, we next examined previously characterized superinfection exclusion systems of Salmonella enterica subsp. anatum phage ε15 and E. coli phage HK97. Salmonella phage ε15 encodes an inhibitor of the cellular α-polymerase (iap) and an O-antigen polymerase (wzyβ) at the 3’ end of its tail operon (Fig. 4e, Extended Data Fig. 2d). These proteins act together to alter the O-antigen structure on the cell surface, changing the natural α-linkages between the sugars to β-linkages³⁸. This serotype conversion, which confers phage resistance on the lysogen, is a distinct mechanism from the O-antigen acetylation mediated by phage JBD44 described above. To determine if this serotype conversion protects the phage progeny from the Kronos effect, we created a deletion mutant that lacked these two genes (ε15Δ^wyzβ-iap) and assessed the effect on phage production. Cells containing an ε15 prophage spontaneously release progeny during growth at 37 °C, with ~5×10⁵ pfu/mL accumulating following overnight incubation (Fig. 4f, left). By contrast, we discovered that the ε15Δ^wyzβ-iap mutant lacking these genes accumulated 100- to 1000-fold fewer progeny in the supernatant than that wild-type phage at the 3- and 6-hour timepoints, and 10⁵-fold less at 20 hours. There was no significant difference in the number of phage progeny produced upon mitomycin C induction (Fig. 4f, right), showing that phage replication and assembly was not affected in the ε15Δ^wyzβ-iap deletion mutant. We next investigated the superinfection exclusion protein in E. coli phage HK97. This protein, known as Gp15, is encoded between the tail assembly chaperone and the tape measure protein (Fig. 4g, Extended Data Fig. 2d). Gp15 localizes to the cell membrane and prevents phage DNA from entering the cytoplasm¹². Deletion of gene 15 from HK97 (HK97^Δ15) led to a 100-fold decrease in spontaneously produced phage progeny following overnight incubation as compared to wild-type HK97 (Fig. 4h, left). As with the other phages, mitomycin induction of the wild type and mutant phage showed no difference in the number of progeny produced (Fig. 4h, right). Together, these data show that the anti-Kronos effect is conserved in temperate phages that infect S. enterica and E. coli.

To investigate how widespread our experimentally confirmed anti-Kronos systems are, we used PSI-BLAST to query the NCBI Virus database and identify homologues of these proteins encoded in complete phage genomes (Extended Data Table 1). We discovered that while Zip is only found in Pseudomonas phages, JBD44 Gp40–41 protein homologues are found in Pseudomonas, Salmonella, Caulobacter, Gordonia, Escherichia and Yersinia phages, and LESϕ3 Gp50 homologues are found in Pseudomonas, Vibrio, Aeromonas, Delftia and Synechococus phages. Similarly, homologues of HK97 Gp15 are found in both E. coli and Klebsiella phages, while the Salmonella Wzyβ-Iap systems are also encoded in phages that infect Pseudomonas. As the five anti-Kronos systems that we identified in this study were all found to be encoded in the phage morphogenetic region (Fig. 4, Extended Data Fig. 2d), we next investigated whether the homologues also co-localize with phage morphogenetic genes. To do this, we retrieved five genes upstream and downstream of each identified homologue and annotated these genes using a collection of HMMs that are specific for phage head and tail genes. We chose the five gene cut-off as it has been previously shown that phage genes within five open reading frames of each other in the genome are frequently co-transferred (i.e. they are genetically linked)³⁹. We discovered that 93% of the putative anti-Kronos genes were found within five open reading frames of a gene or genes that encode phage virion structural genes (Extended Data Table 1). This co-localization makes sense from a biological standpoint as the cell surface binding determinants encoded by the tail proteins need to co-segregate with the anti-Kronos genes to ensure that any recombinant phages maintain their anti-Kronos effect. To more broadly investigate how widespread the anti-Kronos effect may be, we identified previous studies in which adsorption of a given phage to its own lysogen was tested. In addition to the phages that we experimentally validated above, we found evidence of anti-Kronos activity for 29 additional phages. These phages infect a wide range of bacteria, including Pseudomonas, E. coli, Salmonella, Burkholderia, Shigella, Serratia, Proteus, Listeria, Staphylococcus and Streptococcus (Extended Data Table 2). Taken together, these data show that many different phages across a wide range of bacterial species encode anti-Kronos factors.

Discussion

In this work, we have shown that phage Sie systems are not solely expressed for incoming phage defence, but importantly drive viral spread through the bacterial community. It has long been known that prophages express a variety of proteins that provide them with defence against further phage infection^40–42 and mathematical modeling predicted that natural selection would favour the evolution of lysogens that are resistant to their own phage⁴³. Many of the Sie defences have been shown to act at the cell envelope, blocking the cell surface binding or genome injection of superinfecting phages. Here, we characterized a superinfection exclusion protein that binds the PilZ assembly chaperone and modulates P. aeruginosa type IV pilus biosynthesis. The expression of Zip from the JBD26 prophage resulted in the production of shorter pili, which are associated with both decreased twitching motility and increased phage resistance. This resistance is mediated not only against competing phages, but also phages that are spontaneously produced by the JBD26 lysogen. In fact, the anti-Kronos effect mediated by Zip confers a strong advantage to JBD26 lysogens by allowing large numbers of spontaneously produced phage progeny to accumulate in the culture. These phages can then go on to infect other non-lysogenic cells in the community and thereby promote the spread of phage JBD26. We show that this anti-Kronos effect – protecting the lysogen community from loss of phage progeny by reinfection – is mediated by different prophages using different biological mechanisms. These anti-Kronos proteins provide a strong competitive advantage to prophages that encode them by enabling them to efficiently transmit both vertically as lysogens and horizontally via phage infection.

While the precise mechanism through which Zip inhibits phage infection is not yet known, our data allow us to propose a model. A previous study using cryo-electron tomography suggested that only one type of motor, PilB extension or PilT retraction, can bind to the pilus machinery at any given time²⁹. Further work using fluorescence microscopy developed a model where competitive binding of the PilB extension and PilT retraction motors coordinates the repetitive cycles of pilus extension and retraction³⁰. This binding process was proposed to be stochastic, with PilB binding to the inner membrane complex and mediating pilus extension before dissociating and opening the binding site to the PilT retraction motor. Our protein interaction data showed that Zip forms a tripartite complex with PilB/PilZ, and fluorescence microscopy showed that Zip-GFP localizes to the bacterial cell pole, where the pilus is known to assemble. These data suggest that PilB/PilZ/Zip together bind to the pilus assembly site in the inner membrane. We believe that the addition of Zip to this complex destabilizes the interaction of PilB with the pilus assembly complex, causing it to spontaneously dissociate more rapidly than in the absence of Zip. This activity would result in the formation of shorter pili, consistent with what we observed in the JBD26 lysogen. Further support for this model is provided by the dynamics data collected on the JBD26 lysogen, which showed that while the pilus extension velocity was the same as wild-type cells, the pilus extension time was decreased, consistent with the PilB motor prematurely dissociating. Why shorter pili might increase phage resistance is not entirely clear, and this work may hint at a role for the pilus inner membrane complex and/or PilB assembly ATPase in the phage genome entry process. In contrast to E. coli phages, for which several inner membrane proteins are known to be important for phage genome entry^13,44–46, there are no confirmed inner membrane protein requirements for Pseudomonas phages. As PilZ is highly conserved across all P. aeruginosa strains, it provides an excellent target to modulate to provide resistance to a broad spectrum of pilus-dependent phages.

Zip expression is controlled by LasR, the principal regulator of quorum sensing in P. aeruginosa, allowing it to be expressed in a cell density-dependent manner. This permits the lysogen to tune expression in concert with the relative risk of phage infection. At low cell density, there is little risk of being reinfected by phage progeny produced by spontaneous induction from individual bacteria as the cells are spread out. As cell density increases, the chance of infection also increases as the absolute number of cells spontaneously producing phages increases, and the cells within the community are closer together. Thus, it is highly beneficial for the lysogens to respond to this increased risk by modulating expression of their pili and thereby decreasing the chance of phage progeny reinfecting the lysogen. This defence is also protective against other phages that require the type IV pilus for infection that might invade the bacterial community and initiate an epidemic. This quorum-mediated control of anti-phage defence is similar to what has previously been observed for bacterially encoded anti-phage defences. For example, in P. aeruginosa, LasR contributes to high cell density-mediated expression of anti-phage defence systems both directly by upregulating the production of a community-based anti-phage defence molecule known as PQS⁴⁷ and indirectly by upregulating expression of RhlR, which drives expression of CRISPR-Cas systems⁴⁸. While phages have been shown to integrate quorum sensing into their lysis-lysogeny decision⁴⁹, to our knowledge, this work provides the first evidence that prophages have connected into this conserved bacterial circuitry to control expression of their antiviral defence genes. As quorum sensing systems are broadly distributed across bacteria, this is likely a widespread feature of prophage-encoded defences.

Previous studies investigating superinfection exclusion and its effects on viral populations showed that allowing superinfection yields populations that are more capable of adapting to changes in the environment as a result of the potential for exchange of genetic material between viruses^50,51. However, mutants that can prevent superinfection have also been shown to display a significant advantage over their superinfection permissive counterparts, even when this ability came with a substantial cost to their growth rate⁵⁰. Together, these results suggest that while superinfection exclusion is a winning strategy in the short term, preventing superinfection can negatively impact the long-term prospects of a viral population. The quorum mediated control of a superinfection exclusion protein like Zip allows the modulation of activity in a manner that provides anti-phage defence in some environments and permits uptake of genetic material in others, thereby granting both evolutionary benefits. For example, in spatially structured environments, the competition for hosts is local, and lysogens are more likely to be battling genetically identical phages released by nearby cells. In this case there is no advantage to allowing phages to enter the lysogen. Zip expression would be upregulated in direct measure with cell density in the structured environment and exclude these phages from infection, thereby also enabling higher numbers of free phages to escape the structured community in search of new hosts. By contrast, in well-mixed environments where competition is global and lysogens are more likely to encounter genetically diverse phages, low Zip expression could allow other pilus-dependent phages to infect, thereby providing the advantage of genetic exchange. Thus, the measured control of Zip activity provides a means by which prophages can balance these two opposing evolutionary advantages.

The success of a virus depends on its ability to spread both within and between hosts⁵². With bacterial viruses, the spread within host is provided via vertical transmission of the prophage from mother to daughter cells that naturally occurs when cells divide. Spread between hosts involves the horizontal movement of the phage to uninfected cells. For this to efficiently proceed, the phages produced by a lysogenic community need to avoid cells that already contain a prophage so they can efficiently seek out cells in which they can initiate a fruitful infection. Zip activity promotes this by inhibiting the ability of phages to bind to lysogens and reinfect. We call this the anti-Kronos effect. We have shown that this effect has a large impact on free phage numbers produced by lysogens for two other temperate P. aeruginosa phages that are unrelated to JBD26 and use different mechanism to block cell surface binding by phages. In addition, we have shown that anti-Kronos is a conserved feature in temperate phages that infect E. coli and Salmonella, and we expect that the anti-Kronos effect will manifest in many other diverse temperate phage systems. This property of viral self-exclusion has also been described in eukaryotic viruses like HIV, where Nef downregulates CD4+, the viral receptor to support viral propagation, and vaccinia virus, where viral proteins were shown to mediate repulsion of virions from the surface of infected cells, thereby speeding up viral spread^9,53,54. Our work shows that this common eukaryotic virus trait is conserved among bacterial viruses, providing another example where the interaction of bacterial viruses and their hosts is mirrored in eukaryotic systems^55–57.

Resource availability

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, Karen Maxwell (karen.maxwell@utoronto.ca).

Methods

Bacterial strains and phages

Detailed list of all phages and strains available in Supplementary Table 2. E. coli NEB 5-alpha (New England Biolabs) was the cloning strain used. E. coli BL21(λDE3) was used for protein expression and purifications, E. coli BTH101 was used for bacterial two-hybrid assays and E. coli 594 was used for phage-related experiments. Pseudomonas aeruginosa strains were obtained from a previous study²¹. Unless otherwise stated the cells were grown in Lysogeny Broth (LB) with or without 10 mM MgSO₄, with or without 0.1% L-arabinose and 1.2 or 1.5% agar. Specific antibiotic concentrations used for E. coli were: ampicillin (100 μg/mL), tetracyline (12 μg/mL), gentamicin (20 μg/mL) and streptomycin (34 μg/mL) and for P. aeruginosa: carbenicillin (300 μg/mL), gentamicin (50 μg/mL) and tetracycline (90 μg/mL). Strains generated for this study and phages used are detailed in the Key Resources table.

Identification of tfpZ within the IATS O3 genome

A member of the type V pilus family was identified by analyzing the genomic region between nadC and pilB of the 20 IATS strains using RAST SEEDViewer server⁶⁸, looking for the presence of pilA and tfpZ. Polymerase chain reaction (PCR) using primers specific to tfpZ were used to confirm presence of the gene.

Twitching assay

Individual colonies grown on LB agar supplemented with 50 μg/ml gentamicin and 0.1% L-arabinose were stab inoculated into a 1% LB agar plate supplemented with 50 μg/ml gentamicin and 0.1% arabinose and were incubated for 24 h at 37 °C. Twitching motility was assessed by removing the agar and staining the cell mass adhered to the petri plate using 1% crystal violet. Twitching assays with wild-type P. aeruginosa and the lysogens were performed in the absence of antibiotics.

Phage preparations

High titer phage lysates were obtained by growing P. aeruginosa lysogens overnight in LB medium at 37 °C. The cells were diluted 1/100 into fresh LB medium and grown with shaking at 37 °C to an OD₆₀₀ of 0.5. Mitomycin C was added to a final concentration of 3 μg/ml, and cultures were incubated until they lysed. Cellular debris was removed by centrifugation at 15,000 x g for 10 min. The supernatant containing the phages was transferred into a fresh test tube and a few drops of chloroform were added for storage at 4 °C. Phage titers were determined by plating ten-fold serial dilutions on plates containing 150 μL bacteria resuspended in 3 ml of 0.7 % top agar supplemented with 10 mM MgSO₄ (plus relevant antibiotic and 0.1 % L-arabinose when noted). Plates were incubated overnight at 30 °C to allow plaques to form.

Generating mutant phages in P. aeruginosa and Salmonella anatum

To create Zip knockouts in JBD26 and JBD24 a conserved sequence within zip was chosen to generate a guide RNA specific to the 1C CRISPR-CAS system. The guide RNA was cloned into pHERD30T and transformed into the PAO1^1C and PAO1^1CΔcas3 strains⁶⁰. To generate the mutations, lysates of JBD26 and JBD24 were isolated from lysogens within PAO1 grown overnight. PAO1^1CΔcas3 expressing the guide RNA targeting zip was grown at 37 °C overnight in the presence of 50 μg/ml of gentamicin, 0.1 % L-arabinose and 0.5 mM IPTG. The following morning the culture was diluted 1/100 in the same media and was grown at 37 °C until mid-exponential stage. Ten microliters of JBD26 or JBD24 were added to the culture, and it was incubated overnight at 37 °C. The following morning supernatants containing the replicated phages were plated on lawns of PAO1^1C expressing the guide RNA targeting zip with the LB plates and top agar supplemented with the same selection and inducing agents. Following overnight incubation, individual plaques were resuspended in LB and three plaque purification steps were performed. Mutant phages were confirmed by sequencing.

To delete the proposed serotype converting genes 40 and 41 from phage JBD44, a deletion construct was generated to remove the region between nucleotides 28672–31193. We created a construct in pEXG2⁶⁶ that had 250 base pairs flanking either end of this region to use for homologous recombination. This construct encoded the first 15 amino acids of Gp40 and the final 15 amino acids of Gp41. This construct was used for homologous recombination. Phage mutants were confirmed by DNA sequencing.

To remove the serotype converting genes of ε15 in S. anatum, a repair template was generated in pHERD30T to delete nucleotides 23291–24639. We synthesized guide RNAs specific to wzyβ and cloned them into a pCas9 plasmid⁶⁷. S. anatum ε15 lysogens were transformed by the repair template. The transformants were grown overnight at 37 °C in the presence of 10 μg/ml gentamicin and the following morning 2 μl cell-free lysates were plated on lawns of S. anatum expressing pCas9. Individual plaques were isolated, and plaque purified three times in the presence of pCas9 to generate pure phage lysates. The phage genome mutations were confirmed by DNA sequencing.

Phage production over time

Three independent P. aeruginosa lysogens were grown overnight at 30 °C with shaking. The following morning bacteria from 100 μl aliquots of the cell cultures were collected by centrifugation and washed three times with 1 mL of LB to remove all free phages. Equal numbers of cells were resuspended in a final volume of 2 mL LB supplemented with 10 mM MgSO₄ and were incubated at 30 °C with shaking. Samples were taken at 3, 6, 9, 12, and 20 hours; the bacterial cells were collected by centrifugation and serial dilutions were plated onto LB agar and incubated at 30 °C overnight to enumerate the number of bacteria present at each time point, and the titers of the phages present in the supernatant fraction were determined through plating assays on PA14. The LESϕ3 assays were performed with lysogens and lawns of PAO1. The S. anatum lysogens were treated as above with the following amendments: the cultures and lawns were grown at 37 °C and 50 μl of S. anatum resuspension was used for the initial subculture. To prepare the lawns of S. anatum, 100 μl of exponentially growing cells were added to 3 ml of 0.5% top agar lacking MgSO₄. The HK97 lysogens were treated the same as the P. aeruginosa lysogens, except that they were incubated at 37 °C instead of 30 °C.

P. aeruginosa adsorption assays

Overnight cultures of wild-type P. aeruginosa and relevant lysogens were grown at 37 °C with shaking. The following morning the cultures were diluted 1/100 into fresh LB supplemented with 10 mM MgSO₄ and grown to an OD₆₀₀ of 0.8. Cells were collected by centrifugation and resuspended in fresh media three times to remove any free phages from the supernatant. Phages were added at an MOI of 0.1 and the cultures were incubated at 37 °C for 5 minutes. Cells and adsorbed phages were collected by centrifugation and serial dilutions of the supernatant were plated on lawns of sensitive cells to quantify the number of free phages remaining.

Zip promoter assay

The intergenic region between gene 60 and zip was amplified from phage JBD26 and cloned into the NcoI/HindIII sites of the promoterless β-galactosidase reporter shuttle vector pQF50. To analyze promoter type 2, the intergenic region between genes 57 and 58 was amplified from phage JBD24 and cloned into the NcoI/HindIII sites of pQF50. The constructs were transformed into the designated strains and assayed for β-galactosidase activity. Individual colonies were used to inoculate overnight cultures of LB containing 300 μg/ml of carbenicillin. The following morning cells were diluted into fresh LB with carbenicillin and grown at 37 °C to an OD₆₀₀ of 0.5 for exponential phase experiments, or 16 h to reach saturation (OD₆₀₀ was measured at this point). β-Galactosidase activity was quantified by mixing 100 μL of culture with 900 μL of Z buffer (0.06 M Na₂HPO₄, 0.04 M NaH₂PO₄-H₂O, 0.01 M KCl, 0.001 M MgSO₄ and 0.05 M β-mercaptoethanol), the cells were lysed with 100 μl of 1% SDS and 100 μl chloroform and vortexed for 10 sec. The cultures were incubated at 30 °C for 5 min and 200 μl of 40 mg/ml o-nitrophenyl-β-galactosidase resuspended in Z buffer was added before incubating for 15–30 min. The reaction was stopped with the addition of 500 μl 1M Na₂CO₃, A₄₂₀ and A₅₅₀ were measured, and the Miller Units were calculated.

Zip cellular localization

To determine the cellular localization of Zip at native expression levels, a fusion construct of the zip promoter sequence and free-use green fluorescent protein (fuGFP) fused to the N-terminus of zip was cloned into pHERD30T. Twitching inhibition and phage resistance of cells expressing prom-fuGFP-Zip were measured. In tandem fuGFP alone was cloned into the pHERD30T construct. The cultures were inoculated on plates containing 50 μg/ml of gentamicin and 0.1% L-arabinose and were grown overnight at 37 °C. The following day a single colony of each strain was resuspended in 10 μl of 1X PBS and 2 μl was spotted on 2% agarose pads. For all microscopy, Zip localization was assessed using differential interference contrast (DIC) microscopy or using the enhanced green fluorescent protein (EGFP) channel on a Zeiss Axio Imager M1 microscope (Carl Zeiss) at the same exposure time. Three independent fields were captured for each sample and images are representative of three technical replicates across two biological replicates.

Pilus dynamics experiments

Pilus labeling, data acquisition, and data analysis were performed as described previously²⁸. Briefly, cells were grown overnight and used directly in high cell density experiments or diluted 1:1000 and grown to mid-log phase (OD = 0.3) for low cell density experiments. Cells were incubated with 35 ng/μl Alexa Fluor 488 maleimide dye for 45 minutes to stain pili. Unbound dye was washed away by pelleting cells twice and resuspending in fresh medium. Cells were spread on an agarose pad covered with a coverslip for imaging under a Nikon TiE or Nikon Ti2 fluorescent microscope with a 100x NA1.45 Ph3 objective lens. Videos of pilus dynamics were recorded for 30 s each. Pilus dynamics were analyzed in ImageJ⁶⁹.

Bacterial two-hybrid assay

To identify an interaction partner for Zip, we used a library of pilus biosynthesis genes cloned into the Euromedex bacterial two-hybrid system²³. The E. coli BTH101 strains lacking a functional adenylate cyclase were transformed by pKT25-zip and individual pilus genes cloned into pUT18C. The resulting transformants were plated on LB agar plates containing ampicillin (100 μg/mg) and kanamycin (50 μg/ml). The plates were incubated at 30 °C overnight and three individual colonies were selected per plate to assay for interactions. LB cultures were grown overnight at 30 °C in the presence of the stated ampicillin and kanamycin concentrations and supplemented with 0.5 mM IPTG. The next morning 2 μl of culture was spotted onto two sets of chromogenic plates: LB agar with X-gal (40 μg/ml) and MacConkey agar, both of which contained ampicillin (100 μg/ml), kanamycin (50 μg/ml) and 0.5 mM IPTG. The plates were incubated for 24 h at 30 °C. The presence of a blue colony (X-gal) or a red colony (acidification of MacConkey medium) indicates a positive interaction. The overnight culture was then diluted 1/100 in fresh supplemented media and grown at 30 °C with shaking for 6 h. The Miller units were obtained by assaying for β-galactosidase activity as described above.

Protein expression and purification

Single expression of His₆-Zip was achieved used p15TV-L (Genbank: EF456736) and PilZ was cloned in site 1 of pETDuet in frame with the N-terminal hexa-histidine (6His) tag. For co-expression, genes were cloned into pETDuet with Zip in MCS-1 in-frame with the N-terminal 6-His tag and PilZ in MCS-2. Plasmids were propagated in E. coli DH5a supplemented with 100 μg/ml of ampicillin. To express three proteins at once, PilB was cloned into pCDF1-b using the NcoI and SacI restriction sites to remove the 6His-tag. Co-expressions with pETDuet His₆-Zip-PilZ and pCDF-PilB were supplemented with 100 μg/ml of ampicillin and 34 μg/ml streptomycin. Purifications were performed from 2 L cultures of E. coli BL21(λDE3) as outlined below.

The cells were grown at 37 °C with shaking until the OD₆₀₀ reached ~0.8, at which point the culture was induced with 1 mM IPTG and the temperature decreased to 20 °C for expression overnight. The following morning cells were harvested by centrifugation and resuspended in 200 ml of binding buffer: Tris-HCl pH 8.0 with 250 mM NaCl. The cells were lysed by sonication and cellular debris was removed by centrifugation at 20,000 x g for 20 min. The lysate was incubated with Ni-NTA resin and washed with binding buffer containing 30 mM imidazole and eluted with 250 mM imidazole. The elution fractions were separated on 15% Tris-Tricine gels.

Bioinformatic analyses

To explore the diversity and conservation of Sie proteins identified in this work, we performed exhaustive PSI-BLAST searches against the NCBI Virus Genomes Database⁷⁰, which contained 14,000 complete phage genomes (December 2024). To investigate the genomic context of the diverse protein homologues identified, we retrieved five proteins upstream and downstream of the target proteins. Using a large set of HMMs that correspond to the various functions associated with phage morphogenetic proteins, we annotated these neighboring open reading frames and then manually examined the annotations to determine whether each of the phage Sie protein homologues reside in phage morphogenetic regions.

Extended Data

Extended Data Table 1.

Summary of the diversity of previously characterized superinfection exclusion systems and their co-localization with morphogenetic genes. Complete dataset presented in Supplementary Data 1.

Protein	Query	Function	Phage homologs species	# homologues	# homologues associated with morphogenetic region	% homologues associated with morphogenetic region (%)
Zip	YP_010299255.1	Reduces type IV pilus length and piliation	Pseudomonas	87	87	100
Gp50	WIL01470.1	Prevents adsorption of type IV pilus dependent phages	Pseudomonas, Synechococcus, Vibrio, Aeromonas, Delftia	18	16	89
Gp40–41	YP_009275527.1, YP_009275528.1	Adds O-acetyl group to O-antigen repeats of lipopolysaccharide	Pseudomonas, Caulobacter, Salmonella, Gordonia, Escherichia, Yersinia	98	82	84
Wzyβ-Iap	QDB70935.1	Alters the intra-molecular linkage of O-antigen repeats of LPS	Salmonella Pseudomonas	12	11	92
Gp15	NP_037709.1	Inner membrane protein that prevents DNA injection	Escherichia, Klebsiella	11	11	100
GtrABC	AKJ74362.1 AKJ74363.1 AKJ74364.1 AKJ74365.1 AKJ74366.1	Prevents adsorption by adding a side-chain glucose to O-antigen repeats of LPS	Enterobacteria, Salmonella, Shigella, Hafnia	48	38	79
Gp20	QDB70935.1	Adds O-acetyl group to O-antigen repeats of lipopolysaccharide	Pseudomonas, Escherichia, Salmonella, Yersinia, Mycobacterium, Gordonia, Acidianus	92	76	83
TarP	NP_835519.1	Modifies cell wall techoic acid	Actinobacter, Arthrobacter, Bacillus, Gordonia, Microbacteriu, Mycobacterium, Staphylococcus, Streptomyces	91	44	48
Cor	YP_007151630.1	Binds FhuA and blocks phage binding	Erwinia, Escherichia, Pseudomonas, Salmonella, Shigella, Yersinia	356	353	99

Extended Data Table 2.

Studies where lysogens were shown to inhibit adsorption of their cognate phage.

Phage	Bacterial Species	Reference
D3	Pseudomonas aeruginosa	1
vB_Pae_QDWS	Pseudomonas aeruginosa	2
φ80	Escherichia coli	3
N15	Escherichia coli	4
hf4	Escherichia coli	5
ΦV10	Escherichia coli	6
A3 and A4	Salmonella enterica	7
ε34	Salmonella enterica	8
P22	Salmonella enterica	9
BTP1	Salmonella enterica	10
ε34	Salmonella enterica	8
φE125	Burkholderia thailandensis	11
SfII	Shigella flexneri	12
SfIV	Shigella flexneri	12
Sf6	Shigella flexneri	13
X	Serratia marcescens	14
12/57	Proteus mirabilis	15
34/13	Proteus mirabilis	15
36/13	Proteus mirabilis	15
24/25/14	Proteus mirabilis	15
6/24/6	Proteus mirabilis	15
12/57/57	Proteus mirabilis	15
14/20/14	Proteus mirabilis	15
14/20/20	Proteus mirabilis	15
49/F7	Proteus vulgaris	15
50/F7	Proteus vulgaris	15
A188	Listeria monocytogenes	16
ΦHSIC	Listeria pellagia	17
LPP-1	Phormidium uncinatum	18
ΦN315	Staphylococcus aureus	19
TP-J34	Streptococcus thermophilus	20

^1.Newton, G. J. et al. Three-component-mediated serotype conversion in Pseudomonas aeruginosa by bacteriophage D3. Mol Microbiol 39, 1237–1247 (2001).

^2.Xuan Guanhua, Lin Hong, & Wang Jingxue. Expression of a Phage-Encoded Gp21 Protein Protects Pseudomonas aeruginosa against Phage Infection. Journal of Virology 96, e01769–21 (2022).

^3.Kozyrev, D., Svarchevskiĭ, A., Zaĭtsev, E. & Rybchin, V. [Lysogenic conversion induced by phages phi 80. I. A description of the phenomenon and the cloning of the conversion gene]. Genetika 18, 555–560 (1982).

^4.Vostrov A A, Vostrukhina O A, Svarchevsky A N, & Rybchin V N. Proteins responsible for lysogenic conversion caused by coliphages N15 and phi80 are highly homologous. Journal of Bacteriology 178, 1484–1486 (1996).

^5.Kulikov Eugene E. et al. Equine Intestinal O-Seroconverting Temperate Coliphage Hf4s: Genomic and Biological Characterization. Applied and Environmental Microbiology 87, e01124–21 (2021).

^6.Perry, L. L. et al. Sequence analysis of Escherichia coli O157:H7 bacteriophage ΦV10 and identification of a phage-encoded immunity protein that modifies the O157 antigen. FEMS Microbiology Letters 292, 182–186 (2009).

^7.Wollin R, Stocker B A, & Lindberg A A. Lysogenic conversion of Salmonella typhimurium bacteriophages A3 and A4 consists of O-acetylation of rhamnose of the repeating unit of the O-antigenic polysaccharide chain. Journal of Bacteriology 169, 1003–1009 (1987).

^8.Villafane, R., Zayas, M., Gilcrease, E. B., Kropinski, A. M. & Casjens, S. R. Genomic analysis of bacteriophage ε34 of Salmonella entericaserovar Anatum (15+). BMC Microbiology 8, 227 (2008).

^9.Kim, M. & Ryu, S. Spontaneous and transient defence against bacteriophage by phase-variable glucosylation of O-antigen in almonella enterica serovar Typhimurium. Molecular Microbiology 86, 411–425 (2012).

^10.Kintz, E. et al. A BTP1 prophage gene present in invasive non-typhoidal almonella determines composition and length of the O-antigen of the lipopolysaccharide. Molecular Microbiology 96, 263–275 (2015).

^11.Woods Donald E., Jeddeloh Jeffrey A., Fritz David L., & DeShazer David. Burkholderia thailandensis E125 Harbors a Temperate Bacteriophage Specific for Burkholderia mallei. Journal of Bacteriology 184, 4003–4017 (2002).

^12.Allison, G. E. & Verma, N. K. Serotype-converting bacteriophages and O-antigen modification in Shigella flexneri. Trends in Microbiology 8, 17–23 (2000).

^13.Clark, C. A., Beltrame, J. & Manning, P. A. The oac gene encoding a lipopolysaccharide O-antigen acetylase maps adjacent to the integrase-encoding gene on the genome of Shigella flexneri bacteriophage Sf6. Gene 107, 43–52 (1991).

^14.Steiger, H., Müller, U. & Bauer, G. Non-receptivity for ϰ phage of ϰ-lysogenic Serratia and reactions to superinfection of receptive cells with a mutant prophage. Molecular and General Genetics MGG 114, 358–367 (1972).

^15.COETZEE, J. N. Lysogenic Conversion in the Genus Proteus. Nature 189, 946–947 (1961).

^16.Bielmann, R. et al. Receptor binding proteins of Listeria monocytogenes bacteriophages A118 and P35 recognize serovar-specific teichoic acids. Virology 477, 110–118 (2015).

^17.Williamson S. J., McLaughlin M. R., & Paul J. H. Interaction of the ΦHSIC Virus with Its Host: Lysogeny or Pseudolysogeny? Applied and Environmental Microbiology 67, 1682–1688 (2001).

^18.Bisen, P. S., Bagchi, S. N. & Audholia, S. Nitrate reductase activity of a cyanobacteriumPhormidium uncinatum after cyanophage LPP-1 infection. FEMS Microbiology Letters 33, 69–72 (1986).

^19.Ingmer Hanne, Gerlach David, & Wolz Christiane. Temperate Phages of Staphylococcus aureus. Microbiology Spectrum 7, 10.1128/microbiolspec.gpp3-0058-2018 (2019).

^20.Sun, X., Göhler, A., Heller, K. J. & Neve, H. The ltp gene of temperate Streptococcus thermophilus phage TP-J34 confers superinfection exclusion to Streptococcus thermophilus and Lactococcus lactis. Virology 350, 146–157 (2006).

Extended Data Fig. 1 |.

(a) Bacterial Two Hybrid assay querying potential Zip interaction with proteins involved in P. aeruginosa pilus assembly. (b) Twitching motility assays for PAO1, JBD26 lysogen, JBD26Δzip lysogen, and PAO1 expressing Zip-GFP from a plasmid. Three independent biological replicates are shown.

Extended Data Fig. 2 |.

(a) DNA sequences of the predicted promoter regions of the zip genes found in phages JBD26, JBD24, and PA8. The LasR binding motif as predicted by the PePPER webserver is shown as a red box. (b) Protein sequence alignment of Zip homologues found in complete phage genomes. A sequence identity cut-off of 95% was used to remove redundant sequences. (c) Genomic context of zip homologues including near the phage tail operon, while zip^PA8 is encoded at the other end of the morphogenetic region, small terminase (ST) large terminase (LT). (d) Broader genetic context of the anti-Kronos factors from Figure 4. Head decorator (HD), Head tail joining (HTJ), tail assembly chaperone (TAC) tail terminator (TT).

Extended Data Fig. 3 |.

(a) Genome alignment of the phages used in the main study and the adsorption assay. (b) Adsorption assay in P. aeruginosa of designated phages against i. wild-type ii. lysogens.

Data and code availability

This paper does not report original code.

Any information required to reanalyze the data reported in this paper is available from the lead contact upon request.