Biofilm spatial structure and superinfection immunity modulate inter-phage competition

James B. Winans; Carey D. Nadell

doi:10.1371/journal.pbio.3003737

Abstract

Obligately lytic (virulent) phages always lyse host cells to release progeny viruses, while temperate phages can either lyse their hosts or integrate into host genomes as prophages, forming lysogens. There is a rich history of work studying the relative advantages and disadvantages of these two phage life history strategies, but little of this work has addressed the spatial constraints common to biofilm environments. We developed a live imaging system to track lytic infections, lysogenic infections, and uninfected cells at single-cell resolution within three-dimensional Escherichia coli biofilms. We find that biofilm structure substantially impacts the ecological success of different phage infection strategies. Temperate phages have the unique capacity to release phages from lysogens that have undergone lytic induction from within the interior of mature biofilms. When this occurs in biofilm contexts that do not limit phage diffusion, lytic infections expand rapidly, but lysogenic infections are favored as phage mobility declines in densely packed biofilm architectures. In matrix-replete biofilms that do limit phage mobility, lytic phage infection is more limited, favoring lysogenic growth. Direct competition assays between lysogenized host bacteria and obligately lytic phages—with or without the ability to superinfect lysogens—revealed that spatial structure and superinfection potential together greatly impact phage competition outcomes during co-invasion into pre-existing, phage-susceptible biofilm populations. Highly packed, phage diffusion-impeding biofilms disproportionately favored temperate phages in the lysogenic cycle over obligate lytic phages, highlighting how biofilm architecture can constrain lytic phage infection and promote vertical phage genome transmission strategies.

Citation: Winans JB, Nadell CD (2026) Biofilm spatial structure and superinfection immunity modulate inter-phage competition. PLoS Biol 24(3): e3003737. https://doi.org/10.1371/journal.pbio.3003737

Academic Editor: Jeremy J. Barr, Monash University, AUSTRALIA

Received: September 9, 2025; Accepted: March 18, 2026; Published: March 31, 2026

Copyright: © 2026 Winans, Nadell. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data is available within the manuscript and Supporting information files.

Funding: C.D.N. received support from NIGMS grant 1R35GM151158-01 and the Simons Foundation (award number 826672). J.B.W. received support from the NIH T32 AI007519 and 1F31AI183523-01A1. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Abbreviation: PDMS, poly-dimethylsiloxane

Introduction

Temperate and obligately lytic (virulent) phages follow distinct life history strategies [1,2]. Lytic phages inject their genome into a host cell, commandeer their hosts’ replication machinery, and rapidly lyse the cell to release progeny virions [3,4]. Temperate phages can choose to follow the same lytic pathway, or alternatively they can integrate their genome into that of the host as a prophage, converting the cell into a lysogen [5–7]. In both cases, the phage co-opts host resources to propagate, but while lytic infections result in imminent cell death, prophages are often less intrusive to their hosts’ replication, transmitting the phage genome through vertical inheritance until reentry into the lytic cycle [8–10]. Classic theory and more recent work have explored the fitness tradeoffs incurred by these distinct phage strategies, examining how host availability, multiplicity of infection, and environmental conditions influence the success of lysis versus lysogeny in competition [11–18]. However, experimental work that addresses competition for host cells between lytic and temperate phages has focused primarily on bulk liquid culture, rather than on surface-attached groups (biofilms) that bacteria often occupy [18–22]. Recent studies have illustrated that when host bacteria secrete extracellular matrix and grow into biofilm cell clusters, they can collectively slow or halt phage diffusion amongst them [23–31].

Previously, we explored in detail how temperate phage λcI₈₅₇ interacts with biofilm-dwelling host populations of E. coli, finding that if host biofilm matrix production and corresponding architecture are established, then λcI₈₅₇ virions can infect hosts along the biofilm exterior but are unable to access hosts on the biofilm interior [24]. As a consequence, λcI₈₅₇ lysogens arise along the biofilm periphery, where they are predisposed to disperse back into the surrounding environment, which itself can lead to over-representation of λcI₈₅₇ lysogens in new biofilms generated in downstream locations [24]. Here we build on this work by interrogating whether phages spontaneously released by lysogens from within the biofilm interior can mobilize to infect new hosts effectively and thereby alter resident biofilm structure over time. Furthermore, we aimed to assess how λcI₈₅₇ lysogens and virulent λ derivatives might compete with one another in the process of invading a biofilm of susceptible hosts with variable matrix secretion. We approached this question with the simple hypothesis that host genome-embedded prophages and phage virions would have different capabilities to spread when released within biofilms, and also different abilities to invade pre-existing biofilm populations. If supported, this would indicate that the spatial constraints and host physiology specific to biofilm environments are important to consider for the evolutionary dynamics of temperate versus virulent bacteriophage strategies.

We developed an experimental system allowing for live imaging of λ phage particles, lytically infected cells, lysogenized cells, and uninfected host cells, all differentiated by distinct fluorescent reporters within biofilm communities grown under continuous flow [24]. Using this system, we explore two general scenarios under which temperate phages λ might interact differently with host cells in the biofilm context relative to virulent phages of otherwise similar infection characteristics (λvir and λΔcI). First, temperate phages have the distinguishing ability to release phages from within the interior of biofilms in which lysogens have become embedded. We test here how effectively phage virions can spread from biofilm-embedded lysogens undergoing lytic activation, and how their propagation depends on host biofilm architecture. Second, lysogenized host cells and phage virions—due to differences in size, surface properties, and presence/absence of active motility—may differ in their ability to invade biofilms from their exterior, which in turn may also depend on biofilm architecture. To directly compete the two phage lifestyles, we co-invade lysogenized cells and lytically locked phages that either could or could not successfully superinfect lysogens (λvir and λΔcI, respectively) into biofilm communities. By introducing isogenic phages with these different infection properties into biofilms of varying architecture and tracking their spatial propagation over time, we explore how phage life history strategy, superinfection potential, and spatial structure collectively shape the competitive ability of distinct phage life histories in the biofilm context.

Results

Biofilm structure modulates spread of phages released from lysogens embedded within biofilms

A potential advantage distinct to temperate phages in the biofilm context is the ability to disseminate from lysogens present early during biofilm growth and embedded in the interior of the host population [24]. To investigate under what conditions this potential advantage occurs, we established lysogen-embedded E. coli biofilms in curli-producing (WT AR3110, denoted curli⁺) or curli non-producing (ΔcsgBA, denoted curli⁻) E. coli strain backgrounds. In prior work, while testing for pleiotropic effects of the ΔcsgBA deletion, we did not detect any physiological differences between the parental AR3110 strain and its ΔcsgBA derivative with respect to λ phage plaquing ability, adsorption rate, or phage titer population dynamics in well-mixed liquid culture [24]. In biofilm environments, secreted curli polymer matrix, and the multicellular architecture associated with it, greatly reduces phages’ ability to diffuse into biofilms from their exterior [23–25]; however, it is less clear how phages might disseminate if released from lysogens that co-colonized a surface with uninfected cells, leading to mixed-strain populations with lysogens present on the interior.

We first grew E. coli biofilms comprised of uninfected E. coli and E. coli lysogenized with λcI₈₅₇ at an inoculation ratio of 10:1. In one set of experiments, both uninfected and infected cells were curli⁺; in a second set of experiments, both strains were curli⁻. Our primary model λ phage harbors the temperature-sensitive cI₈₅₇ allele, such that lysogens are relatively stable and spontaneously induce to the lytic cycle at a rate of 10⁻⁵ in our room temperature biofilm culture conditions (S1 Fig). Because the total bacterial count in our culture chambers was on the order of 10⁶, spontaneous lytic induction by λcI₈₅₇ lysogens and subsequent phage spread did not occur frequently enough to be detected without experimental intervention [24]. After 72 h of biofilm growth, we therefore induced some but not all lysogenized cells to enter the lytic cycle via a short heat treatment, after which cultures were returned to room temperature (see Methods). This resulted in 1% and 7% lytic induction in the curli⁺ and curli⁻ biofilm contexts, respectively (S2 Fig). We then tracked these systems with daily imaging for 120 h (Fig 1A–1C). To assess population dynamics and compare the relative abilities of phages versus lysogens to replicate—and in keeping with recent theory [32,33]—we quantify lysogens and virocells (i.e., host cells undergoing productive lytic phage infection) as the primary fitness metrics. The absolute number of lysogens in these experiments initially declined following lytic induction and then increased in abundance thereafter (Fig 1B).

Download:

Fig 1. λcI₈₅₇ lytic infections and host lysogenization following induction of lysogens embedded within Escherichia coli biofilms with either WT curli production (curli⁺) or no curli production (curli⁻).

Host biofilms were inoculated with λcI₈₅₇ lysogens and uninfected host cells at a 1:10 ratio and cultured for 72 h prior to λcI₈₅₇ phage induction. (A) Time series of biofilms inoculated with lysogens and uninfected E. coli hosts that either produce (top row) or do not produce (bottom row) curli extracellular matrix. The first image in each row was taken just after heat was applied to induce some biofilm-embedded lysogens to switch to the lytic infection cycle. (B, C) Population dynamics of active lytic infections, lysogens, and uninfected host cells in curli⁺ and curli⁻ biofilms (n = 7–24). (D) Population dynamics of infected cells that are currently virocells (in the lytic pathway) in both structured and unstructured biofilms (n = 7–24). (F, G) Time-resolved cell packing frequency distributions within curli⁺ and curli⁻ biofilms. For panels B–D, center points denote medians, and bars denote interquartile ranges. All figure panels are new analyses derived from data published in Winans and colleagues (2025) [24]. The numerical data within this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.g001

In biofilms of the AR3110 (curli⁺) background, the susceptible host population was mostly protected from phage exposure and maintained net positive growth (Fig 1B). Lytic infections were visible at low frequency; at maximum, they represented less than 5% of the total cell population, and among all infected cells (virocells and lysogens), lytic infections peaked at ~15% by 96 h and then later declined again (Fig 1D). The results indicate that temperate phage virions can indeed spread when released within the interior of curli-producing E. coli biofilms, albeit to a limited extent. This observation is consistent with previous reports, which showed that growing bacterial populations can support a lytic phage population, dependent on spatial constraints [23,25,34]. In curli⁻ biofilms, on the other hand, virocell frequencies rose within 24 h after λcI₈₅₇ induction to ~88% of the total infected cell population (Fig 1C and 1D). Thereafter, as lysogen counts increased sharply, virocell biovolume consistently declined due to the depletion of available uninfected, phage-sensitive hosts. Lysogen counts surpassed uninfected host cells by 48 h and ultimately comprised the majority of the bacterial population (Fig 1A and 1C).

As noted above, our primary model phage λ carries the cI₈₅₇ allele, which removes cI Repressor’s wild type sensitivity to host RecA and instead confers temperature sensitivity to cI; this allowed us to control the timing of lytic induction via heat treatment. To ensure that the results above were representative of spontaneous lytic induction by true wild-type λ phages with cI RecA sensitivity, we reconstructed the wild-type RecA-sensitive allele in the parental λ strain, such that it still contained the fluorescent protein constructs required for visualizing host lysogenization and phage virions (this strain is denoted λcI_WT). We then repeated the experiments depicted in Fig 1B and 1C and found the same qualitative population dynamics for phages, lysogens, and uninfected host E. coli when embedded in both AR3110 (curli⁺) and ΔcsgBA (curli⁻) host populations. The main difference between the results for λcI_WT and those shown in Fig 1B and 1C was the timing of phage propagation and lysogen spread in the case of the ΔcsgBA host biofilm context (S3 Fig).

The two distinct patterns of phage spread in AR3110 curli⁺ versus ΔcsgBA curli⁻ populations imposed correspondingly distinct feedback on the architecture of the biofilms in which they were occurring. Because phage-mediated killing and subsequent lysogen regrowth are expected to directly restructure local cellular crowding, we measured cell packing frequency distributions at 1 -day intervals following the heat induction step of these experiments. In the AR3110 curli⁺ E. coli background, the frequency distribution of cell packing changed negligibly over the course of induction and subsequent phage spread, reflecting the limitation on phage diffusion and new lytic infections imposed by curli-replete biofilm architecture (Fig 1E). In the curli⁻ background, by contrast, the whole population was down-shifted for cell packing density distribution due to the large fraction of cells infected and killed by phages released from induced lysogens. After most of the curli⁻ population was lysogenized, the biofilms then recovered to their original cell packing distributions (Fig 1F).

We were curious as to whether the results above could be due to reduced expression of the λ phage receptor (LamB maltoporin) by E. coli in the AR3110 curli⁺ population, which reaches larger absolute size and cell packing than the ΔcsgBA curli⁻ background. To control for this possibility, we produced an sfGFP transcriptional reporter for the mal operon (see Methods). When maltose is supplied as the sole carbon source, as was the case for all main text experiments, the mal operon reporter was active in the large majority of the biofilm population, regardless of spatial location. Variation in phage receptor expression therefore cannot explain the greatly reduced rate of lytic and lysogenic infection in curli⁺ experiments (S4 Fig).

Altogether, these results demonstrate that temperate phages can indeed propagate after induction of biofilm-embedded lysogens, but the fraction of host cells accessible to released phage virions is constrained by the presence of extracellular matrix and the packing architecture that matrix confers to host cells. In curli⁺ biofilms, impeded diffusion of phage particles limits widespread infection, and sensitive cells remain in the majority among host E. coli. In curli⁻ biofilms, where phages can diffuse more freely, phages released from induced lysogens cause a rapid increase in new infections throughout the population. While phages in the lytic cycle temporarily outnumber phages in the lysogenic cycle, a decrease in sensitive cells and lysogenic cell accumulation leads to a drop in virocell frequency over time and longer-term dominance of the lysogen sub-population.

Biofilm structure and superinfection immunity shape temperate and virulent phage propagation during invasion into pre-established biofilms

We next investigated how temperate and virulent λ phages compete during co-invasion into pre-formed E. coli biofilms. In particular, we compared the biofilm invasion abilities of temperate phage λcI₈₅₇ lysogens and λ phages genetically restricted to the lytic cycle, including λvir (able to infect lysogens) and λΔcI (unable to infect lysogens). Both phages can adsorb to host bacteria lysogenized by λ, but λΔcI is suppressed following genome injection [35,36]. The rationale for co-invading λcI₈₅₇ lysogens with virulent λ phages was to explore the potential advantages and disadvantages of being embedded within a bacterial host—which in principle allows temperate phages to take advantage of native host physiology that they do not themselves encode—when invading a pre-existing biofilm population. Depending on resident biofilm architecture, other bacteria may or may not be able to invade and take hold within the resident population, as has been explored previously [26,37,38]. Phage particles also display variable biofilm invasion and infection potential as a function of phage virions’ biophysical properties and host biofilm architecture [24,25,28,39]. Co-invasion of resident biofilms by both lysogenized cells and virulent phages adds complexity to the invasibility question, as both the invading phages and lysogens can potentially change the composition and architecture of the resident biofilm environment. Whether or not virulent phages can superinfect co-invading lysogens is another important parameter to explore for this question, and to accommodate it we performed all of the experiments below (in separate treatments) with either superinfecting virulent phages (λvir) or non-superinfecting virulent phages (λΔcI)

In separate experiments, we performed phage and lysogen co-invasions into E. coli biofilms formed by strains with no curli production (ΔcsgBA, denoted curli⁻), wild type curli production (AR3110 parental strain, denoted curli⁺), or curli overexpression (csgD*, denoted curli⁺⁺); the three host strains display no significant difference in their efficiency of plating for any of the λ phages used here (S5 Fig). We grew biofilms of these strains for 72 h prior to introducing equal titers of virulent phages and λcI₈₅₇ lysogens (10⁴ [PFU/CFU]µL⁻¹ for 2 h), after which biofilms were imaged daily for 120 h to quantify phage and lysogen proliferation, as well as uninfected host population dynamics (S6 Fig). The 72 h cultivation period prior to phage/lysogen invasion was chosen because it allows for biofilm matrix production to a steady state for these culture conditions. Additionally, the biofilms produced by E. coli in this time frame do not become large enough for metabolic stratification to occur via growth substrate depletion on the biofilm interior [40–42]. This mitigates the concern that bacteria on the biofilm interior might be less susceptible to infection due to stalled metabolic activity.

The introduced lysogens were in the AR3110 (curli⁺) background in all cases for consistency across treatments, and we confirmed in earlier work that introducing a minority population of curli⁺ cells does not substantially alter resident biofilm architecture on the time scale of these experiments [27,39]. To control for the possibility of phage-lysogen interaction within the syringes during the 2 h invasion period, we performed separate experiments in which the invasion inoculum mixtures (λcI₈₅₇ lysogens with λvir phages; or λcI₈₅₇ lysogens with λΔcI phages) were sampled every 15 min for 2 h; these controls revealed no substantial changes in lysogen or phage titer over the 2 h time period (S7 Fig).

The central motivating questions for these experiments were (1) whether it is more favorable under direct competition to invade a pre-existing biofilm as a virulent phage particle or as a prophage inside a host, and (2) whether the answers to the first question change as a function of host biofilm architecture and/or superinfection exclusion status of the lysogens.

We observed phage invasion into all 3 E. coli biofilm variants (curli⁻, curli⁺, curli⁺⁺), but to greatly varying degrees and with different relative success of lysogens and phage virions depending on both biofilm architecture and lysogen superinfection exclusion (Fig 2A). Consistent with prior reports [23,25], the addition of lytic phages to the curli⁻ strain caused the biomass of the resident biofilm to decrease, while curli⁺ and curli⁺⁺ biofilms maintained net positive growth after phage addition (Fig 2B and 2C). To assess competition between λcI₈₅₇ prophages and virulent phages, we focused on the sub-population of host E. coli that were lysogenized or undergoing lytic cycle infection. This sub-set of bacterial host cells represents the resource pool that was available to phages over the course of the experiments: we monitored the virocell infected fraction, indicating lytic phage reproduction, and the lysogenized fraction, indicating temperate prophage reproduction. Note again that in these experiments, the spontaneous lytic induction rate of temperate λcI₈₅₇ is below detection, so we are effectively asking whether phage genomes embedded within host cells compete better during biofilm invasion than the virion particles of virulent λ derivatives, which must pass from one lytic host infection to another in order to propagate. We used this approach to determine whether virulent or temperate phage strains best competed for host access, dependent on the cell packing architecture of host biofilms as well as the superinfection exclusion status of lysogenized hosts.

Download:

Fig 2. Invasion of resident, λ-sensitive biofilms with λcI₈₅₇ lysogens and virulent derivative phages that either could not superinfect lysogenized hosts (λΔcI) or could superinfect lysogenized hosts (λvir).

(A) Representative images of biofilms at 120 h after invasion of non-superinfecting phages (top row) or superinfecting phages (bottom row) and lysogens into curli⁻ (left column), curli⁺ (middle column), or curli⁺⁺ (right column) biofilms. (B) Proportional change in uninfected E. coli population sizes for curli⁻, curli⁺, and curli⁺⁺ biofilms at the end of the 120 h experimental monitoring period (n = 12–14). P-values denote the outcome of Wilcoxon tests against a null prediction of 1.0, which would indicate no net change in uninfected host cells over the course of the experiments. (C) Absolute quantification of uninfected E. coli at 120 h for each of the three matrix-producing strain variants. P-values denote the outcome of Mann-Whitney U tests for pairwise comparisons against curli⁺, which represents wild type curli regulation (n = 12–14). (D) Frequency of lysogens among phage-infected cells when co-invaded into curli⁻, curli⁺, or curli⁺⁺ biofilms with virulent λΔcI phages that cannot superinfect lysogens. P-values denote the outcomes of Wilcoxon tests against a median value of 0.5, which would indicate neutral competition between λcI₈₅₇ lysogens and λΔcI virocell infections (n = 11–22). (E) Frequency of lysogens among phage-infected cells when co-invaded into curli⁻, curli⁺, or curli⁺⁺ biofilms with virulent λvir phages that can superinfect lysogens. P-values denote the outcomes of Wilcoxon tests against a median value of 0.5 (n = 6–16). The numerical data within this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.g002

When λcI₈₅₇ lysogens and λΔcI phages were invaded together into curli⁻ or curli⁺ biofilms, outcomes were variable, but on average the λcI₈₅₇ lysogens and λΔcI virocells were both present among phage-infected hosts (Fig 2A, top row and 2D). By contrast, λcI₈₅₇ lysogens represented nearly 100% of actively replicating phage genomes in the majority of cases when invading matrix-overexpressing csgD* (curli⁺⁺) populations; λΔcI virocells were sometimes present but always rare, indicating minimal host accessibility (Fig 2D). When λcI₈₅₇ lysogens and λvir phages were invaded together, the phage-host population dynamics changed, especially when host biofilms were ΔcsgBA (curli⁻), allowing for free phage diffusion (Fig 2A, bottom row and 2E). When the dual invasion was performed into curli⁻ populations, we observed very few lysogens and a large majority of lytic λvir infections by the end of the experiment (Fig 2E). This result follows from the simple logic that when host biofilm structure poses no barrier to phage diffusion, and virulent phages can super-infect lysogens, the majority of E. coli (lysogenic or not) will be killed by lytic infections. There was no systematic difference in the relative replication success of λcI₈₅₇ lysogens and λvir infections when the two were co-invaded into AR3110 (curli⁺) E. coli host biofilms, similar to what we observed for co-invasion of λcI₈₅₇ lysogens and λΔcI phages (Fig 2E). Likewise, when invading curli⁺⁺ matrix-overexpressing host biofilms, λcI₈₅₇ lysogens constituted nearly 100% of the phage-infected host cells in all runs of the experiment.

Discussion

We aimed in this study to assess how temperate lysogens and virulent phages might differ in their abilities to interact with host bacteria in biofilm environments, particularly when introduced together to the same biofilm population of susceptible hosts. Temperate phages have the capacity to burst from a lysogenized host that is already inside the biofilm interior; we showed here that this capacity translates to new host infections and lysogenization events, contingent on whether host biofilm structure impedes phage diffusion. Further, direct competition between temperate lysogens and virulent phages during co-invasion into biofilms has distinct outcomes that also depend on the extent of phage diffusion limitation, as well as lysogens’ superinfection exclusion status. In matrix-deficient ΔcsgBA (curli⁻) biofilms—which do not substantially impede phage diffusion—temperate lysogens with superinfection exclusion and virulent phages were able to colonize and propagate in comparable quantities. On the other hand, lysogens were always outcompeted by superinfection-capable virulent phages in curli⁻ biofilm settings. On the time scales assessed here, interestingly, AR3110 (curli⁺) biofilms generally supported coexistence of lysogens and virulent phages, both superinfecting and non-superinfecting. Finally, matrix over-expressing csgB* (curli⁺⁺) biofilms strongly favored host-embedded temperate lysogens over virulent phages, regardless of superinfection ability.

A central motivation for this work was the notion that temperate prophages that are replicated along with host cell division also passively benefit from the other host features—for example, dedicated structures for motility and surface attachment—that might be particularly helpful for surviving in biofilm environments. However, another key consideration is that lysogens often require more time for nutrient uptake to grow and divide, whereas virulent phages may replicate more rapidly while depending on access to susceptible hosts for successive rounds of lytic infections. Our results support the intuition that when phage mobility is high and lysogens are not immune to superinfection, virulent phage strategies are particularly successful during biofilm invasion. By contrast, when hosts are tightly packed together within biofilms and phage mobility is low, temperate lysogens are relatively more effective at invading uninfected host populations, regardless of superinfection exclusion. Biofilm populations with intermediate or locally variable phage mobility allowed for invasion of both virulent phages and temperate lysogens. These outcomes are all consistent with the original idea that biofilm structure can alter the relative fitness of virulent versus temperate phages, but future study will be required for establishing what role biofilm-type environments might play more generally in the evolutionary dynamics of temperate versus lytic phage life history strategies. Allowing for repeated rounds of biofilm growth, phage exposure, host and phage dispersal, exit from and re-entry into lysogeny by temperate phages, and variation in the amount of time phages and hosts spend in well-mixed planktonic versus biofilm conditions are all likely to be important for understanding how and why biofilm growth might be important to consider for phage life history evolution.

A key distinction between temperate and virulent phages lies in temperate phages’ ability to actively regulate whether they propagate via vertical transmission through lysogenized host replication or horizontal transmission through lytic infection events, choosing between lysis and lysogeny in response to environmental cues [43–45]. However, in our study, the direct competition experiments involved phages that were locked into either the lytic or lysogenic pathway, a deliberate choice here to clarify the outcome of the experiments. Future research can build on these findings with experiments on longer time scales that allow for temperate phages to switch back and forth between lytic and lysogenic propagation—all while in direct competition with virulent phages that only replicate via lytic infection. Our work here offers a start to this research direction by demonstrating that spatial structure and superinfection immunity can shift the outcome of competition between lysogens and obligately lytic phages.

Methods

Strain construction

The E. coli strains used in this study were are derivatives of AR3110, which was generated from the K-12 parental strain W3110 by restoration of its biofilm matrix regulation relative to the domesticated ancestor [46]. λvir was generated by introducing three previously identified point mutations that lock λ into the lytic pathway upon host infection and render it insensitive to silencing by λ prophages (i.e., bypassing superinfection immunity) [36]. λΔcI was made via deletion of the cI locus encoding Repressor, locking this phage strain into the lytic pathway as well, but without bypassing prophage superinfection immunity. λcI_WT was constructed by heat-inducing a strain lysogenized with λcI₈₅₇ to activate the Red genes prior to transforming in a DNA fragment that encodes cI_WT [47]. Transformed cells were rescued in 5 mL of SOC media overnight at 30 °C. Then 500 µL of the rescued cells were back-diluted into 30 mL of LB media and incubated at 30 °C for 3 h prior to treatment with Mitomycin C at a concentration of 2.5 µgmL for 3 h. Mitomycin C treatment selected for prophages that are sensitive to the host SOS response and temperature-insensitive. This culture was then centrifuged to pellet lysogens, and the supernatant was filter sterilized to remove any remaining bacterial cells. Phages in suspension were plated for plaque formation. Resultant plaques were then screened for cI_WT by sequencing. The malK reporter plasmid was produced by placing the malK promoter region upstream of sfGFP, and transformed into designated strain backgrounds. All strains constructed for this study are listed in Table 1 and will be provided upon request.

Download:

Table 1. Bacterial strains, phage strains, and software packages used in the study.

https://doi.org/10.1371/journal.pbio.3003737.t001

Microfluidic device fabrication

Microfluidic flow chambers were fabricated by casting poly-dimethylsiloxane (PDMS, Dow SYLGARD 184) onto prepared device molds (schematic provided in S8 Fig). PDMS molds were trimmed, inlet/outlet ports punched, and then bonded to plasma-cleaned #1.5 glass coverslips (Azer Scientific, cat. #1152260). Internal chamber dimensions were 5,000 μm × 500 μm × 70 μm (L × W × H). Media flow connections were established using inlet tubing (Cole Parmer PTFE #30) connected via 27-gauge needles (BD Precision) to 1 mL syringes driven by Harvard Apparatus syringe pumps. Outlet tubing was taped to a waste petri dish to collect chamber effluent.

Biofilm culture conditions

Non-lysogenic E. coli AR3110 strains were grown overnight at 37 °C with shaking in LB media; strains lysogenized with λcI₈₅₇ were cultured overnight at 30 °C due to temperature-sensitive repressor proteins. Overnight cultures were normalized to OD₆₀₀ = 1.0 before chamber inoculation. After an initial 45-min static incubation to facilitate bacterial attachment, chambers were perfused continuously with M9 minimal medium supplemented with 0.5% maltose at 0.1 μL/min. To test for mal operon expression, strains holding the malK-sfGFP plasmid were grown in either M9 with 0.5% glucose or M9 with 0.5% maltose, and sfGFP fluorescence was measured as a function of cell packing in both media conditions. All biofilm experiments were performed at room temperature.

Lytic induction and spontaneous induction assays

To measure phage burst size, lysogenic E. coli was plated for CFU counts before and after lytic induction, and the total amount of phages produced by one round of induction was measured by PFU plating. To measure spontaneous induction of lysogenic strains within biofilms, monocultures of λcI_WT AR3110 and λcI₈₅₇ AR3110 lysogens were incubated for 72 h and then measured for frequency of lysogens producing the mTurquoise phage capsid label. This assay was performed in monocultures, so that phages produced by spontaneous induction did not lead to further phage infection and replication. For temperate phage induction within biofilms, microfluidic devices containing susceptible host E. coli and E. coli lysogenized with λcI₈₅₇ (10:1 ratio) were incubated for 72 h prior to lytic induction. The Repressor protein of λcI₈₅₇ variant used in this study is temperature-sensitive. To induce some of the lysogens into the lytic cycle, we incubated microfluidic devices at 42 °C for 40 min, which results in ~1%–10% of the lysogen population converting to lytic cycle replication. Following heat treatment, biofilms were monitored for 120 h to quantify the population dynamics of uninfected susceptible cells, lysogenized cells, and virocells undergoing lytic infection. For comparison with lysogens carrying the λcI_WT prophage, we repeated this co-culture experiment with susceptible host E. coli and E. coli lysogenized with λcI_WT and monitored the cocultures daily for 168 h without any heat treatment or other experimental manipulation.

Phage propagation and efficiency of plating

λcI₈₅₇ phages were generated from lysogenized E. coli via heat shock at 42 °C, followed by incubation at 37 °C until host cell lysis. λcI_WT phages were prepared by treating lysogenized E. coli with 2.5 µgmL of mitomycin C followed by incubation until host cell lysis. Obligately lytic phages (λΔcI and λvir) were generated by inoculating 250 µL of E. coli cultured overnight in λ broth with a −80 °C glycerol strain stock, incubating for 30 min prior to adding 4.75 mL of fresh λ broth to allow for bacterial growth and phage infection and replication to occur. Phage titers were determined by standard plaque assays and back-diluted to 10⁴ PFU/μL in M9 maltose media for experiments. To test whether the different bacterial strains varied in their susceptibility to infection outside of the biofilm condition, we measured the efficiency of plating for λcI₈₅₇, λΔcI, and λvir on ΔcsgBA (curli⁻), AR3110 (curli⁺), and csgD* (curli⁺⁺) host strain backgrounds. This was performed by inoculating monoculture lawns of each strain in soft agar, followed by inoculation of a stock phage dilution series incubating overnight for plaque formation.

Coinvasion of phages and lysogens into established biofilms

Sensitive biofilms of different strain backgrounds (ΔcsgBA [curli⁻], AR3110 [curli⁺], and csgD* [curli⁺⁺]) were grown for 72 h prior to addition of equal titers of phages (either λΔcI or λvir) and lysogens (AR3110 curli⁺) (10⁴ PFU/µL and 10⁴ CFU/µL, respectively) for 2 h. We then reconnected the microfluidic flow cells to sterile M9 media and incubated for 120 h, imaging every 24 h to measure community composition over time. To measure whether the titers of invading phages and lysogens were changing over the course of the coinvasion, λcI₈₅₇ lysogens were incubated with either λvir or λΔcI and the CFU and PFU titers were sampled every 15 min for 2 h.

Microscopy and image analysis

All imaging was done using a Zeiss 980 line-scanning confocal microscope with a 40×/1.2 N.A. water objective. The mTurquoise2 protein fused to the capsids of λcI_WT, λcI₈₅₇, λvir, and λΔcI phages was excited with a 458 laser line (in separate experiments). The sfGFP protein expressed by the malK promoter was excited with a 488 laser line. The mKO2 protein that lysogenized E. coli expresses was excited with a 543 laser line. The mKate2 that E. coli expresses was excited with a 594 laser line. One image was taken to capture the dynamics for each biological replicate per chamber. Prior to export, images were processed by constrained iterative deconvolution in ZEN blue. Images were then imported into the BiofilmQ framework. Constitutive reporter signals were binarized using Otsu thresholding with a manual sensitivity parameter. To remove phages in free virion particle form from virocell biovolume calculations, voxel clusters that were smaller than 0.4µm³ were removed from the analysis. For all analyses, a 3-dimensional grid was used to divide the segmented biovolumes into pseudo-cell cubes that measured 0.72 µm per side. Cell packing measurements combined the biovolume of all bacteria in a sample and calculated the fraction of total biovolume located within 6 μm of each segmented grid cube.

Replication and statistical analysis

Biological replicates are indicated within figure legends. Biological replicates were drawn from distinct microfluidic chambers inoculated independently. Statistical significance was evaluated using Mann–Whitney U tests with Bonferroni corrections for multiple comparisons or one-sample Wilcoxon test against a null comparison value, indicated in figure legends. We chose nonparametric tests as the assumptions underlying parametric tests could not be assessed for our data. Data points within each figure represent independent biological replicates, and trend lines indicate median values. For time course plots, median value with interquartile ranges were plotted.

Supporting information

S1 Fig.

(A) Population burst size of λcI₈₅₇ phages with mTurquoise2 capsid label, which is the fluorescent label used for all experiments (n = 4). (B) Spontaneous induction within biofilm monocultures of AR3110 curli⁺ E. coli lysogenized with λcI₈₅₇ or λcI_WT (Mann–Whitney U-test, n = 4, 8). The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.s001

(PDF)

S2 Fig. Virocell biovolume and frequency immediately after lytic induction within lysogenized E. coli AR3110 (curli⁺) or ΔcsgBA (curli⁻).

(A) Absolute volume of virocells (i.e., cells undergoing active lytic infection) following heat treatment of λcI₈₅₇ lysogens (Mann–Whitney U-test, n = 14–16). (B) Frequency of lytically inducing cells following heat treatment (Mann–Whitney U-test, n = 14–16). (C) Representative image of an AR3110 (curli⁺) biofilm following lytic induction. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.s002

(PDF)

S3 Fig. Quantification of uninfected biovolume, lysogenized biovolume, and virocell biovolume (i.e., cells undergoing active lytic infection) in E. coli biofilms inoculated with non-lysogenized E. coli and lysogenized E. coli at an initial ratio of 10:1.

Here, lysogens carried prophages of λcI_WT, which induces the lytic cycle when its cI repressor is cleaved by host RecA. This is by contrast with experiments in Fig 1 of the main text, which were performed with lysogens carrying prophages of λcI₈₅₇, which induces lytic infection upon a temperature shift to 42 °C. These experiments were performed as controls to document whether the dynamical patterns of phage infection and host lysogenization that were observed in the original experiments with λcI₈₅₇ would be recapitulated with λcI_WT. Here, instead of using heat to induce lytic propagation of prophages inoculated with uninfected hosts, biofilms were grown without disturbance, and λcI_WT prophages induced lytic infection spontaneously without experimental manipulation. As for Fig 1 of the main text, the experiment was performed in (A) an AR3110 (curli⁺) E. coli strain background, and in (B) a ΔcsgBA (curli⁻) strain background. The same qualitative population dynamics were observed here as for Fig 1B and 1C, albeit with delayed lytic propagation and lysogenization of new hosts in the ΔcsgBA (curli⁻) treatment. We speculate that this is due to a lower overall degree of spontaneous lytic induction among λcI_WT lysogens in comparison with the lytic induction of λcI₈₅₇ via the heat treatment manipulation for the experiments depicted in Fig 1 of the main text. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.s003

(PDF)

S4 Fig. Transcription of malK was active regardless of cell position or neighborhood cell packing within biofilms if maltose was the sole carbon source provided.

(A) malK transcriptional reporter intensity (AU) within AR3110 curli⁺ biofilms grown in glucose or maltose, plotted as a function of cell packing (n = 4). (B) Representative image of E. coli grown in maltose, where the GFP-transcriptional malK reporter is active throughout the biofilm. (C) Representative image of E. coli grown in glucose, where the GFP-transcriptional malK reporter is inactive. (D) malK transcription (AU) as a function of cell packing in biofilms of the ΔcsgBA curli⁻ strain in glucose or maltose (n = 4), or (E) in biofilms of the csgD* (curli⁺⁺) strain (n = 5) in glucose or maltose. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.s004

(PDF)

S5 Fig. Comparison of efficacy of phage infection in the three strains used across all experiments in the study—ΔcsgBA (curli⁻); AR3110 (curli⁺), and csgD* (curli⁺⁺)—for each of the main phage variants used in the study—λcI₈₅₇, λΔcI, and λvir.

For each phage strain, no significant differences were observed in efficiency of plating across the three host bacterial different strains used for experiements in this study (n = 3, 4, Mann–Whitney U tests with bonferroni correction). The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.s005

(PDF)

S6 Fig. Illustrative diagram for the design of experiments in Fig 2 of the main text.

In separate experiments, E. coli with varying degrees of curli extracellular matrix production were inoculated and grown for 72 h in biofilm monoculture before being invaded with a mixture of lysogens and phage virions. The E. coli strain backgrounds included the double deletion mutant ΔcsgBA (phenotype denoted curli⁻, shown in red), which cannot produce curli matrix protein; the parental E. coli strain AR3110 (phenotype denoted curli⁺, shown in purple), which produces curli matrix, and finally the csg promoter mutant csgD* (phenotype denoted curli⁺⁺, shown in pink), which produces curli earlier and at higher rates than the parental AR3110 strain. Once grown, these separate biofilm growth chambers were invaded with a mixture of lysogens containing a λcI₈₅₇ prophage (shown in yellow) and virulent λ phage virions (small blue viruses) from one of two different phage strains: λΔcI, which cannot superinfect lysogens, and λvir, which can superinfect lysogenized E. coli. After coinvasion, biofilms were tracked daily for 120 h for changes in uninfected and infected biovolume.

https://doi.org/10.1371/journal.pbio.3003737.s006

(PDF)

S7 Fig. Population dynamics of λcI₈₅₇ lysogens and either (A) virulent phages that cannot superinfect lysogens (λΔcI) or (B) virulent phages that can superinfect lysogens (λvir) in static culture over 2 h (n = 3).

These experiments served as a control to determine whether the ratio of phages and lysogens could have changed within the syringes during the co-culture invasion experiments, whose results are shown in Fig 2 of the main text. On this 2 h time scale, no subtantial differences were observed between the compositions of the mixtures of lysogens with virulent non-superinfecting λΔcI phages versus virulent and superinfecting λvir phages. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003737.s007

(PDF)

S8 Fig. Schematic of the microfluidic devices.

This diagram illustrates 4 parallel chambers that are connected to inflow and outflow tubes through which media is driven at a constant rate. The vacuum port leads to a peripheral channel and is connected to a laboratory vacuum line to minimize introduction of air bubbles into the biofilm chambers.

https://doi.org/10.1371/journal.pbio.3003737.s008

(PDF)

S1 Data. Underlying data.

https://doi.org/10.1371/journal.pbio.3003737.s009

(XLSX)

Acknowledgments

We are grateful to members of the Nadell research group and the Dartmouth microbiology community for their feedback on this project.

References

1. Fillol-Salom A, et al. The secret life (cycle) of temperate bacteriophages. 2021.
2. Howard-Varona C, Hargreaves KR, Abedon ST, Sullivan MB. Lysogeny in nature: mechanisms, impact and ecology of temperate phages. ISME J. 2017;11(7):1511–20. pmid:28291233
- View Article
- PubMed/NCBI
- Google Scholar
3. Koskella B, Hernandez CA, Wheatley RM. Understanding the impacts of bacteriophage viruses: from laboratory evolution to natural ecosystems. Annu Rev Virol. 2022;9(1):57–78. pmid:35584889
- View Article
- PubMed/NCBI
- Google Scholar
4. Pires DP, Melo LDR, Azeredo J. Understanding the complex phage-host interactions in biofilm communities. Annu Rev Virol. 2021;8(1):73–94. pmid:34186004
- View Article
- PubMed/NCBI
- Google Scholar
5. Casjens SR, Hendrix RW. Bacteriophage lambda: early pioneer and still relevant. Virology. 2015;479–480:310–30. pmid:25742714
- View Article
- PubMed/NCBI
- Google Scholar
6. Nguyen TVP, Wu Y, Yao T, Trinh JT, Zeng L, Chemla YR, et al. Coinfecting phages impede each other’s entry into the cell. Curr Biol. 2024;34(13):2841-2853.e18. pmid:38878771
- View Article
- PubMed/NCBI
- Google Scholar
7. Trinh JT, Székely T, Shao Q, Balázsi G, Zeng L. Cell fate decisions emerge as phages cooperate or compete inside their host. Nat Commun. 2017;8:14341. pmid:28165024
- View Article
- PubMed/NCBI
- Google Scholar
8. Pattenden T, Eagles C, Wahl LM. Host life-history traits influence the distribution of prophages and the genes they carry. Philos Trans R Soc Lond B Biol Sci. 2022;377(1842):20200465. pmid:34839698
- View Article
- PubMed/NCBI
- Google Scholar
9. Nepal R, Houtak G, Wormald P-J, Psaltis AJ, Vreugde S. Prophage: a crucial catalyst in infectious disease modulation. Lancet Microbe. 2022;3(3):e162–3. pmid:35544071
- View Article
- PubMed/NCBI
- Google Scholar
10. Bondy-Denomy J, Davidson AR. When a virus is not a parasite: the beneficial effects of prophages on bacterial fitness. J Microbiol. 2014;52(3):235–42. pmid:24585054
- View Article
- PubMed/NCBI
- Google Scholar
11. Chao L, Levin BR, Stewart FM. A complex community in a simple habitat: an experimental study with bacteria and phage. Ecology. 1977;58(2):369–78.
- View Article
- Google Scholar
12. Levin BR, Stewart FM, Chao L. Resource-limited growth, competition, and predation: a model and experimental studies with bacteria and bacteriophage. Am Nat. 1977;111:3–24.
- View Article
- Google Scholar
13. Lenski RE. Coevolution of bacteria and phage: are there endless cycles of bacterial defenses and phage counterdefenses?. J Theor Biol. 1984;108(3):319–25. pmid:6748694
- View Article
- PubMed/NCBI
- Google Scholar
14. Weitz JS. Quantitative viral ecology: dynamics of viruses and their microbial hosts. Princeton University Press; 2016.
15. Weitz JS, Li G, Gulbudak H, Cortez MH, Whitaker RJ. Viral invasion fitness across a continuum from lysis to latency. Virus Evol. 2019;5(1):vez006. pmid:31024737
- View Article
- PubMed/NCBI
- Google Scholar
16. Shivam S, Li G, Lucia-Sanz A, Weitz JS. Timescales modulate optimal lysis-lysogeny decision switches and near-term phage reproduction. Virus Evol. 2022;8(1):veac037. pmid:35615104
- View Article
- PubMed/NCBI
- Google Scholar
17. Marchi J, Khalek C, George A, Weitz JS, Chait R. Stable coexistence and transport of lytic phage infections with migrating bacterial hosts. bioRxiv. 2025.
- View Article
- Google Scholar
18. Kimchi O, Meir Y, Wingreen NS. Lytic and temperate phage naturally coexist in a dynamic population model. ISME J. 2024;18:wrae093.
- View Article
- Google Scholar
19. Rendueles O, de Sousa JAM, Rocha EPC. Competition between lysogenic and sensitive bacteria is determined by the fitness costs of the different emerging phage-resistance strategies. Elife. 2023;12:e83479. pmid:36975200
- View Article
- PubMed/NCBI
- Google Scholar
20. Yan J, Bassler BL. Surviving as a community: antibiotic tolerance and persistence in bacterial biofilms. Cell Host Microbe. 2019;26(1):15–21. pmid:31295420
- View Article
- PubMed/NCBI
- Google Scholar
21. Flemming H-C, van Hullebusch ED, Neu TR, Nielsen PH, Seviour T, Stoodley P, et al. The biofilm matrix: multitasking in a shared space. Nat Rev Microbiol. 2023;21(2):70–86. pmid:36127518
- View Article
- PubMed/NCBI
- Google Scholar
22. Sauer K, Stoodley P, Goeres DM, Hall-Stoodley L, Burmølle M, Stewart PS, et al. The biofilm life cycle: expanding the conceptual model of biofilm formation. Nat Rev Microbiol. 2022;20(10):608–20. pmid:35922483
- View Article
- PubMed/NCBI
- Google Scholar
23. Vidakovic L, Singh PK, Hartmann R, Nadell CD, Drescher K. Dynamic biofilm architecture confers individual and collective mechanisms of viral protection. Nat Microbiol. 2018;3(1):26–31. pmid:29085075
- View Article
- PubMed/NCBI
- Google Scholar
24. Winans JB, Zeng L, Nadell CD. Spatial propagation of temperate phages within and among biofilms. Proc Natl Acad Sci U S A. 2025;122(6):e2417058122. pmid:39903123
- View Article
- PubMed/NCBI
- Google Scholar
25. Winans JB, Wucher BR, Nadell CD. Multispecies biofilm architecture determines bacterial exposure to phages. PLoS Biol. 2022;20(12):e3001913. pmid:36548227
- View Article
- PubMed/NCBI
- Google Scholar
26. Wucher BR, Elsayed M, Adelman JS, Kadouri DE, Nadell CD. Bacterial predation transforms the landscape and community assembly of biofilms. Curr Biol. 2021;31(12):2643-2651.e3. pmid:33826904
- View Article
- PubMed/NCBI
- Google Scholar
27. Bond MC, Vidakovic L, Singh PK, Drescher K, Nadell CD. Matrix-trapped viruses can prevent invasion of bacterial biofilms by colonizing cells. Elife. 2021;10:e65355. pmid:34240700
- View Article
- PubMed/NCBI
- Google Scholar
28. Díaz-Pascual F, Hartmann R, Lempp M, Vidakovic L, Song B, Jeckel H, et al. Breakdown of Vibrio cholerae biofilm architecture induced by antibiotics disrupts community barrier function. Nat Microbiol. 2019;4(12):2136–45. pmid:31659297
- View Article
- PubMed/NCBI
- Google Scholar
29. Wucher BR, Winans JB, Elsayed M, Kadouri DE, Nadell CD. Breakdown of clonal cooperative architecture in multispecies biofilms and the spatial ecology of predation. Proc Natl Acad Sci U S A. 2023;120(6):e2212650120. pmid:36730197
- View Article
- PubMed/NCBI
- Google Scholar
30. Simmons EL, et al. Biofilm structure promotes coexistence of phage-resistant and phage-susceptible bacteria. mSystems. 2020;5:e00877-19.
- View Article
- Google Scholar
31. Simmons EL, Drescher K, Nadell CD, Bucci V. Phage mobility is a core determinant of phage-bacteria coexistence in biofilms. ISME J. 2018;12(2):531–43. pmid:29125597
- View Article
- PubMed/NCBI
- Google Scholar
32. Li G, Cortez MH, Dushoff J, Weitz JS. When to be temperate: on the fitness benefits of lysis vs. lysogeny. Virus Evol. 2020;6(2):veaa042. pmid:36204422
- View Article
- PubMed/NCBI
- Google Scholar
33. Goel T, Beckett SJ, Weitz JS. Eco-evolutionary dynamics of temperate phages in periodic environments. Preprint. 2024. https://www.biorxiv.org/content/10.1101/2024.07.29.604806v1
34. Eriksen RS, Svenningsen SL, Sneppen K, Mitarai N. A growing microcolony can survive and support persistent propagation of virulent phages. Proc Natl Acad Sci U S A. 2018;115(2):337–42. pmid:29259110
- View Article
- PubMed/NCBI
- Google Scholar
35. Brown S, Mitarai N, Sneppen K. Protection of bacteriophage-sensitive Escherichia coli by lysogens. Proc Natl Acad Sci U S A. 2022;119(14):e2106005119. pmid:35344423
- View Article
- PubMed/NCBI
- Google Scholar
36. Atsumi S, Little JW. Regulatory circuit design and evolution using phage lambda. Genes Dev. 2004;18(17):2086–94. pmid:15342489
- View Article
- PubMed/NCBI
- Google Scholar
37. Nadell CD, Drescher K, Wingreen NS, Bassler BL. Extracellular matrix structure governs invasion resistance in bacterial biofilms. ISME J. 2015;9(8):1700–9. pmid:25603396
- View Article
- PubMed/NCBI
- Google Scholar
38. Rugaie OA, Abdellatif AAH, El-Mokhtar MA, Sabet MA, Abdelfattah A, Alsharidah M, et al. Retardation of bacterial biofilm formation by coating urinary catheters with metal nanoparticle-stabilized polymers. Microorganisms. 2022;10(7):1297. pmid:35889016
- View Article
- PubMed/NCBI
- Google Scholar
39. Winans JB, Delgadillo-Guevara M, Popp PF, Erhardt M, Nadell CD. Bacterial conjugation can restructure biofilms and increase their resilience while constraining host cell dispersal. Curr Biol. 2025;35(24):6126-6136.e3. pmid:41401781
- View Article
- PubMed/NCBI
- Google Scholar
40. Teal TK, Lies DP, Wold BJ, Newman DK. Spatiometabolic stratification of Shewanella oneidensis biofilms. Appl Environ Microbiol. 2006;72:7324–30.
- View Article
- Google Scholar
41. Evans CR, Smiley MK, Asahara Thio S, Wei M, Florek LC, Dayton H, et al. Spatial heterogeneity in biofilm metabolism elicited by local control of phenazine methylation. Proc Natl Acad Sci U S A. 2023;120(43):e2313208120. pmid:37847735
- View Article
- PubMed/NCBI
- Google Scholar
42. Wang T, Shen P, He Y, Zhang Y, Liu J. Spatial transcriptome uncovers rich coordination of metabolism in E. coli K12 biofilm. Nat Chem Biol. 2023;19(8):940–50. pmid:37055614
- View Article
- PubMed/NCBI
- Google Scholar
43. Ptashne M. A genetic switch. 3rd ed. Cold Spring Harbor Laboratory Press; 2004.
44. Silpe JE, Bassler BL. A host-produced quorum-sensing autoinducer controls a phage lysis-lysogeny decision. Cell. 2019;176(1–2):268-280.e13. pmid:30554875
- View Article
- PubMed/NCBI
- Google Scholar
45. Zamora-Caballero S, Chmielowska C, Quiles-Puchalt N, Brady A, Gallego Del Sol F, Mancheño-Bonillo J, et al. Antagonistic interactions between phage and host factors control arbitrium lysis-lysogeny decision. Nat Microbiol. 2024;9(1):161–72. pmid:38177302
- View Article
- PubMed/NCBI
- Google Scholar
46. Serra DO, Hengge R. Bacterial multicellularity: The biology of Escherichia coli building large-scale biofilm communities. Annu Rev Microbiol. 2021;75:269–90.
- View Article
- Google Scholar
47. Thomason L, et al. Recombineering: genetic engineering in bacteria using homologous recombination. Curr Protocols Mol Biol. 2007;78:1.16.1-1.16.24.
- View Article
- Google Scholar
48. Hartmann R, Jeckel H, Jelli E, Singh PK, Vaidya S, Bayer M, et al. Quantitative image analysis of microbial communities with BiofilmQ. Nat Microbiol. 2021;6(2):151–6. pmid:33398098
- View Article
- PubMed/NCBI
- Google Scholar

[ref1] 1. Fillol-Salom A, et al. The secret life (cycle) of temperate bacteriophages. 2021.

[ref2] 2. Howard-Varona C, Hargreaves KR, Abedon ST, Sullivan MB. Lysogeny in nature: mechanisms, impact and ecology of temperate phages. ISME J. 2017;11(7):1511–20. pmid:28291233
View Article
PubMed/NCBI
Google Scholar

[3] View Article

[4] PubMed/NCBI

[5] Google Scholar

[ref3] 3. Koskella B, Hernandez CA, Wheatley RM. Understanding the impacts of bacteriophage viruses: from laboratory evolution to natural ecosystems. Annu Rev Virol. 2022;9(1):57–78. pmid:35584889
View Article
PubMed/NCBI
Google Scholar

[7] View Article

[8] PubMed/NCBI

[9] Google Scholar

[ref4] 4. Pires DP, Melo LDR, Azeredo J. Understanding the complex phage-host interactions in biofilm communities. Annu Rev Virol. 2021;8(1):73–94. pmid:34186004
View Article
PubMed/NCBI
Google Scholar

[11] View Article

[12] PubMed/NCBI

[13] Google Scholar

[ref5] 5. Casjens SR, Hendrix RW. Bacteriophage lambda: early pioneer and still relevant. Virology. 2015;479–480:310–30. pmid:25742714
View Article
PubMed/NCBI
Google Scholar

[15] View Article

[16] PubMed/NCBI

[17] Google Scholar

[ref6] 6. Nguyen TVP, Wu Y, Yao T, Trinh JT, Zeng L, Chemla YR, et al. Coinfecting phages impede each other’s entry into the cell. Curr Biol. 2024;34(13):2841-2853.e18. pmid:38878771
View Article
PubMed/NCBI
Google Scholar

[19] View Article

[20] PubMed/NCBI

[21] Google Scholar

[ref7] 7. Trinh JT, Székely T, Shao Q, Balázsi G, Zeng L. Cell fate decisions emerge as phages cooperate or compete inside their host. Nat Commun. 2017;8:14341. pmid:28165024
View Article
PubMed/NCBI
Google Scholar

[23] View Article

[24] PubMed/NCBI

[25] Google Scholar

[ref8] 8. Pattenden T, Eagles C, Wahl LM. Host life-history traits influence the distribution of prophages and the genes they carry. Philos Trans R Soc Lond B Biol Sci. 2022;377(1842):20200465. pmid:34839698
View Article
PubMed/NCBI
Google Scholar

[27] View Article

[28] PubMed/NCBI

[29] Google Scholar

[ref9] 9. Nepal R, Houtak G, Wormald P-J, Psaltis AJ, Vreugde S. Prophage: a crucial catalyst in infectious disease modulation. Lancet Microbe. 2022;3(3):e162–3. pmid:35544071
View Article
PubMed/NCBI
Google Scholar

[31] View Article

[32] PubMed/NCBI

[33] Google Scholar

[ref10] 10. Bondy-Denomy J, Davidson AR. When a virus is not a parasite: the beneficial effects of prophages on bacterial fitness. J Microbiol. 2014;52(3):235–42. pmid:24585054
View Article
PubMed/NCBI
Google Scholar

[35] View Article

[36] PubMed/NCBI

[37] Google Scholar

[ref11] 11. Chao L, Levin BR, Stewart FM. A complex community in a simple habitat: an experimental study with bacteria and phage. Ecology. 1977;58(2):369–78.
View Article
Google Scholar

[39] View Article

[40] Google Scholar

[ref12] 12. Levin BR, Stewart FM, Chao L. Resource-limited growth, competition, and predation: a model and experimental studies with bacteria and bacteriophage. Am Nat. 1977;111:3–24.
View Article
Google Scholar

[42] View Article

[43] Google Scholar

[ref13] 13. Lenski RE. Coevolution of bacteria and phage: are there endless cycles of bacterial defenses and phage counterdefenses?. J Theor Biol. 1984;108(3):319–25. pmid:6748694
View Article
PubMed/NCBI
Google Scholar

[45] View Article

[46] PubMed/NCBI

[47] Google Scholar

[ref14] 14. Weitz JS. Quantitative viral ecology: dynamics of viruses and their microbial hosts. Princeton University Press; 2016.

[ref15] 15. Weitz JS, Li G, Gulbudak H, Cortez MH, Whitaker RJ. Viral invasion fitness across a continuum from lysis to latency. Virus Evol. 2019;5(1):vez006. pmid:31024737
View Article
PubMed/NCBI
Google Scholar

[50] View Article

[51] PubMed/NCBI

[52] Google Scholar

[ref16] 16. Shivam S, Li G, Lucia-Sanz A, Weitz JS. Timescales modulate optimal lysis-lysogeny decision switches and near-term phage reproduction. Virus Evol. 2022;8(1):veac037. pmid:35615104
View Article
PubMed/NCBI
Google Scholar

[54] View Article

[55] PubMed/NCBI

[56] Google Scholar

[ref17] 17. Marchi J, Khalek C, George A, Weitz JS, Chait R. Stable coexistence and transport of lytic phage infections with migrating bacterial hosts. bioRxiv. 2025.
View Article
Google Scholar

[58] View Article

[59] Google Scholar

[ref18] 18. Kimchi O, Meir Y, Wingreen NS. Lytic and temperate phage naturally coexist in a dynamic population model. ISME J. 2024;18:wrae093.
View Article
Google Scholar

[61] View Article

[62] Google Scholar

[ref19] 19. Rendueles O, de Sousa JAM, Rocha EPC. Competition between lysogenic and sensitive bacteria is determined by the fitness costs of the different emerging phage-resistance strategies. Elife. 2023;12:e83479. pmid:36975200
View Article
PubMed/NCBI
Google Scholar

[64] View Article

[65] PubMed/NCBI

[66] Google Scholar

[ref20] 20. Yan J, Bassler BL. Surviving as a community: antibiotic tolerance and persistence in bacterial biofilms. Cell Host Microbe. 2019;26(1):15–21. pmid:31295420
View Article
PubMed/NCBI
Google Scholar

[68] View Article

[69] PubMed/NCBI

[70] Google Scholar

[ref21] 21. Flemming H-C, van Hullebusch ED, Neu TR, Nielsen PH, Seviour T, Stoodley P, et al. The biofilm matrix: multitasking in a shared space. Nat Rev Microbiol. 2023;21(2):70–86. pmid:36127518
View Article
PubMed/NCBI
Google Scholar

[72] View Article

[73] PubMed/NCBI

[74] Google Scholar

[ref22] 22. Sauer K, Stoodley P, Goeres DM, Hall-Stoodley L, Burmølle M, Stewart PS, et al. The biofilm life cycle: expanding the conceptual model of biofilm formation. Nat Rev Microbiol. 2022;20(10):608–20. pmid:35922483
View Article
PubMed/NCBI
Google Scholar

[76] View Article

[77] PubMed/NCBI

[78] Google Scholar

[ref23] 23. Vidakovic L, Singh PK, Hartmann R, Nadell CD, Drescher K. Dynamic biofilm architecture confers individual and collective mechanisms of viral protection. Nat Microbiol. 2018;3(1):26–31. pmid:29085075
View Article
PubMed/NCBI
Google Scholar

[80] View Article

[81] PubMed/NCBI

[82] Google Scholar

[ref24] 24. Winans JB, Zeng L, Nadell CD. Spatial propagation of temperate phages within and among biofilms. Proc Natl Acad Sci U S A. 2025;122(6):e2417058122. pmid:39903123
View Article
PubMed/NCBI
Google Scholar

[84] View Article

[85] PubMed/NCBI

[86] Google Scholar

[ref25] 25. Winans JB, Wucher BR, Nadell CD. Multispecies biofilm architecture determines bacterial exposure to phages. PLoS Biol. 2022;20(12):e3001913. pmid:36548227
View Article
PubMed/NCBI
Google Scholar

[88] View Article

[89] PubMed/NCBI

[90] Google Scholar

[ref26] 26. Wucher BR, Elsayed M, Adelman JS, Kadouri DE, Nadell CD. Bacterial predation transforms the landscape and community assembly of biofilms. Curr Biol. 2021;31(12):2643-2651.e3. pmid:33826904
View Article
PubMed/NCBI
Google Scholar

[92] View Article

[93] PubMed/NCBI

[94] Google Scholar

[ref27] 27. Bond MC, Vidakovic L, Singh PK, Drescher K, Nadell CD. Matrix-trapped viruses can prevent invasion of bacterial biofilms by colonizing cells. Elife. 2021;10:e65355. pmid:34240700
View Article
PubMed/NCBI
Google Scholar

[96] View Article

[97] PubMed/NCBI

[98] Google Scholar

[ref28] 28. Díaz-Pascual F, Hartmann R, Lempp M, Vidakovic L, Song B, Jeckel H, et al. Breakdown of Vibrio cholerae biofilm architecture induced by antibiotics disrupts community barrier function. Nat Microbiol. 2019;4(12):2136–45. pmid:31659297
View Article
PubMed/NCBI
Google Scholar

[100] View Article

[101] PubMed/NCBI

[102] Google Scholar

[ref29] 29. Wucher BR, Winans JB, Elsayed M, Kadouri DE, Nadell CD. Breakdown of clonal cooperative architecture in multispecies biofilms and the spatial ecology of predation. Proc Natl Acad Sci U S A. 2023;120(6):e2212650120. pmid:36730197
View Article
PubMed/NCBI
Google Scholar

[104] View Article

[105] PubMed/NCBI

[106] Google Scholar

[ref30] 30. Simmons EL, et al. Biofilm structure promotes coexistence of phage-resistant and phage-susceptible bacteria. mSystems. 2020;5:e00877-19.
View Article
Google Scholar

[108] View Article

[109] Google Scholar

[ref31] 31. Simmons EL, Drescher K, Nadell CD, Bucci V. Phage mobility is a core determinant of phage-bacteria coexistence in biofilms. ISME J. 2018;12(2):531–43. pmid:29125597
View Article
PubMed/NCBI
Google Scholar

[111] View Article

[112] PubMed/NCBI

[113] Google Scholar

[ref32] 32. Li G, Cortez MH, Dushoff J, Weitz JS. When to be temperate: on the fitness benefits of lysis vs. lysogeny. Virus Evol. 2020;6(2):veaa042. pmid:36204422
View Article
PubMed/NCBI
Google Scholar

[115] View Article

[116] PubMed/NCBI

[117] Google Scholar

[ref33] 33. Goel T, Beckett SJ, Weitz JS. Eco-evolutionary dynamics of temperate phages in periodic environments. Preprint. 2024. https://www.biorxiv.org/content/10.1101/2024.07.29.604806v1

[ref34] 34. Eriksen RS, Svenningsen SL, Sneppen K, Mitarai N. A growing microcolony can survive and support persistent propagation of virulent phages. Proc Natl Acad Sci U S A. 2018;115(2):337–42. pmid:29259110
View Article
PubMed/NCBI
Google Scholar

[120] View Article

[121] PubMed/NCBI

[122] Google Scholar

[ref35] 35. Brown S, Mitarai N, Sneppen K. Protection of bacteriophage-sensitive Escherichia coli by lysogens. Proc Natl Acad Sci U S A. 2022;119(14):e2106005119. pmid:35344423
View Article
PubMed/NCBI
Google Scholar

[124] View Article

[125] PubMed/NCBI

[126] Google Scholar

[ref36] 36. Atsumi S, Little JW. Regulatory circuit design and evolution using phage lambda. Genes Dev. 2004;18(17):2086–94. pmid:15342489
View Article
PubMed/NCBI
Google Scholar

[128] View Article

[129] PubMed/NCBI

[130] Google Scholar

[ref37] 37. Nadell CD, Drescher K, Wingreen NS, Bassler BL. Extracellular matrix structure governs invasion resistance in bacterial biofilms. ISME J. 2015;9(8):1700–9. pmid:25603396
View Article
PubMed/NCBI
Google Scholar

[132] View Article

[133] PubMed/NCBI

[134] Google Scholar

[ref38] 38. Rugaie OA, Abdellatif AAH, El-Mokhtar MA, Sabet MA, Abdelfattah A, Alsharidah M, et al. Retardation of bacterial biofilm formation by coating urinary catheters with metal nanoparticle-stabilized polymers. Microorganisms. 2022;10(7):1297. pmid:35889016
View Article
PubMed/NCBI
Google Scholar

[136] View Article

[137] PubMed/NCBI

[138] Google Scholar

[ref39] 39. Winans JB, Delgadillo-Guevara M, Popp PF, Erhardt M, Nadell CD. Bacterial conjugation can restructure biofilms and increase their resilience while constraining host cell dispersal. Curr Biol. 2025;35(24):6126-6136.e3. pmid:41401781
View Article
PubMed/NCBI
Google Scholar

[140] View Article

[141] PubMed/NCBI

[142] Google Scholar

[ref40] 40. Teal TK, Lies DP, Wold BJ, Newman DK. Spatiometabolic stratification of Shewanella oneidensis biofilms. Appl Environ Microbiol. 2006;72:7324–30.
View Article
Google Scholar

[144] View Article

[145] Google Scholar

[ref41] 41. Evans CR, Smiley MK, Asahara Thio S, Wei M, Florek LC, Dayton H, et al. Spatial heterogeneity in biofilm metabolism elicited by local control of phenazine methylation. Proc Natl Acad Sci U S A. 2023;120(43):e2313208120. pmid:37847735
View Article
PubMed/NCBI
Google Scholar

[147] View Article

[148] PubMed/NCBI

[149] Google Scholar

[ref42] 42. Wang T, Shen P, He Y, Zhang Y, Liu J. Spatial transcriptome uncovers rich coordination of metabolism in E. coli K12 biofilm. Nat Chem Biol. 2023;19(8):940–50. pmid:37055614
View Article
PubMed/NCBI
Google Scholar

[151] View Article

[152] PubMed/NCBI

[153] Google Scholar

[ref43] 43. Ptashne M. A genetic switch. 3rd ed. Cold Spring Harbor Laboratory Press; 2004.

[ref44] 44. Silpe JE, Bassler BL. A host-produced quorum-sensing autoinducer controls a phage lysis-lysogeny decision. Cell. 2019;176(1–2):268-280.e13. pmid:30554875
View Article
PubMed/NCBI
Google Scholar

[156] View Article

[157] PubMed/NCBI

[158] Google Scholar

[ref45] 45. Zamora-Caballero S, Chmielowska C, Quiles-Puchalt N, Brady A, Gallego Del Sol F, Mancheño-Bonillo J, et al. Antagonistic interactions between phage and host factors control arbitrium lysis-lysogeny decision. Nat Microbiol. 2024;9(1):161–72. pmid:38177302
View Article
PubMed/NCBI
Google Scholar

[160] View Article

[161] PubMed/NCBI

[162] Google Scholar

[ref46] 46. Serra DO, Hengge R. Bacterial multicellularity: The biology of Escherichia coli building large-scale biofilm communities. Annu Rev Microbiol. 2021;75:269–90.
View Article
Google Scholar

[164] View Article

[165] Google Scholar

[ref47] 47. Thomason L, et al. Recombineering: genetic engineering in bacteria using homologous recombination. Curr Protocols Mol Biol. 2007;78:1.16.1-1.16.24.
View Article
Google Scholar

[167] View Article

[168] Google Scholar

[ref48] 48. Hartmann R, Jeckel H, Jelli E, Singh PK, Vaidya S, Bayer M, et al. Quantitative image analysis of microbial communities with BiofilmQ. Nat Microbiol. 2021;6(2):151–6. pmid:33398098
View Article
PubMed/NCBI
Google Scholar

[170] View Article

[171] PubMed/NCBI

[172] Google Scholar

Biofilm spatial structure and superinfection immunity modulate inter-phage competition

Biofilm spatial structure and superinfection immunity modulate inter-phage competition

This is an uncorrected proof.

Figures

Abstract

Introduction

Results

Biofilm structure modulates spread of phages released from lysogens embedded within biofilms

Biofilm structure and superinfection immunity shape temperate and virulent phage propagation during invasion into pre-established biofilms

Discussion

Methods

Strain construction

Microfluidic device fabrication

Biofilm culture conditions

Lytic induction and spontaneous induction assays

Phage propagation and efficiency of plating

Coinvasion of phages and lysogens into established biofilms

Microscopy and image analysis

Replication and statistical analysis

Supporting information

S1 Fig.

S2 Fig. Virocell biovolume and frequency immediately after lytic induction within lysogenized E. coli AR3110 (curli⁺) or ΔcsgBA (curli⁻).

S3 Fig. Quantification of uninfected biovolume, lysogenized biovolume, and virocell biovolume (i.e., cells undergoing active lytic infection) in E. coli biofilms inoculated with non-lysogenized E. coli and lysogenized E. coli at an initial ratio of 10:1.

S4 Fig. Transcription of malK was active regardless of cell position or neighborhood cell packing within biofilms if maltose was the sole carbon source provided.

S5 Fig. Comparison of efficacy of phage infection in the three strains used across all experiments in the study—ΔcsgBA (curli⁻); AR3110 (curli⁺), and csgD* (curli⁺⁺)—for each of the main phage variants used in the study—λcI₈₅₇, λΔcI, and λvir.

S6 Fig. Illustrative diagram for the design of experiments in Fig 2 of the main text.

S7 Fig. Population dynamics of λcI₈₅₇ lysogens and either (A) virulent phages that cannot superinfect lysogens (λΔcI) or (B) virulent phages that can superinfect lysogens (λvir) in static culture over 2 h (n = 3).

S8 Fig. Schematic of the microfluidic devices.

S1 Data. Underlying data.

Acknowledgments

References

This is an uncorrected proof.

Figures

Abstract

Introduction

Results

Biofilm structure modulates spread of phages released from lysogens embedded within biofilms

Biofilm structure and superinfection immunity shape temperate and virulent phage propagation during invasion into pre-established biofilms

Discussion

Methods

Strain construction

Microfluidic device fabrication

Biofilm culture conditions

Lytic induction and spontaneous induction assays

Phage propagation and efficiency of plating

Coinvasion of phages and lysogens into established biofilms

Microscopy and image analysis

Replication and statistical analysis

Supporting information

S1 Fig.

S2 Fig. Virocell biovolume and frequency immediately after lytic induction within lysogenized E. coli AR3110 (curli+) or ΔcsgBA (curli−).

S3 Fig. Quantification of uninfected biovolume, lysogenized biovolume, and virocell biovolume (i.e., cells undergoing active lytic infection) in E. coli biofilms inoculated with non-lysogenized E. coli and lysogenized E. coli at an initial ratio of 10:1.

S4 Fig. Transcription of malK was active regardless of cell position or neighborhood cell packing within biofilms if maltose was the sole carbon source provided.

S5 Fig. Comparison of efficacy of phage infection in the three strains used across all experiments in the study—ΔcsgBA (curli−); AR3110 (curli+), and csgD* (curli++)—for each of the main phage variants used in the study—λcI857, λΔcI, and λvir.

S6 Fig. Illustrative diagram for the design of experiments in Fig 2 of the main text.

S7 Fig. Population dynamics of λcI857 lysogens and either (A) virulent phages that cannot superinfect lysogens (λΔcI) or (B) virulent phages that can superinfect lysogens (λvir) in static culture over 2 h (n = 3).

S8 Fig. Schematic of the microfluidic devices.

S1 Data. Underlying data.

Acknowledgments

References

S2 Fig. Virocell biovolume and frequency immediately after lytic induction within lysogenized E. coli AR3110 (curli⁺) or ΔcsgBA (curli⁻).

S5 Fig. Comparison of efficacy of phage infection in the three strains used across all experiments in the study—ΔcsgBA (curli⁻); AR3110 (curli⁺), and csgD* (curli⁺⁺)—for each of the main phage variants used in the study—λcI₈₅₇, λΔcI, and λvir.

S7 Fig. Population dynamics of λcI₈₅₇ lysogens and either (A) virulent phages that cannot superinfect lysogens (λΔcI) or (B) virulent phages that can superinfect lysogens (λvir) in static culture over 2 h (n = 3).