All bacterial biofilms contain an extracellular matrix rich in filamentous molecules that self-associate, conferring emergent properties to bacteria, including antibiotic tolerance. Pseudomonas aeruginosa is a human pathogen that forms biofilms in diverse infectious settings, where the upregulation of a filamentous bacteriophage Pf4, has been shown to be a key virulence factor that protects bacteria from antibiotics. Here, we modeled biophysical characteristics of biofilm-linked liquid crystalline droplets formed by Pf4, which predicted that sub-stoichiometric phage binders had the ability to disrupt liquid crystals by changing the surface properties of the phage. We tested this prediction by developing nanobodies targeting the outer surface of the Pf4 phage, which disrupted in vitro reconstituted droplets, promoted antibiotic diffusion into bacteria, disrupted P. aeruginosa biofilm formation under a variety of conditions, and abolished antibiotic tolerance of biofilms. The inhibition strategy illustrated in this study could be extended to biofilms of other pathogenic bacteria, where filamentous molecules are pervasive in the extracellular matrix. Furthermore, our findings exemplify how targeting a biophysical mechanism, rather than a defined biochemical target, is a promising avenue for intervention, with the potential of applying this concept to other disease-related contexts.

Theoretical modeling predicts small binders disrupt phage self-association

Classically biologic or synthetic binders have been used to target specific biochemical interactions between proteins and their ligands, such as bacterial adhesins [39–41] or in targeted cancer therapies [42]. In the case of Pf4 liquid crystal-mediated antibiotic tolerance, however, there is no specific biochemical interaction to target as liquid crystalline droplets self-assemble through depletion attraction between Pf4 filaments in crowded environments [25,31]. In such environments, the characteristics of Pf4 droplets are governed by biophysical properties such as the size and shape of depletants (crowding biopolymers rich in the matrix) and Pf4 filaments [32].

We therefore used coarse-grained molecular dynamics simulations to explore how the depletion interaction between Pf4 filaments could be suppressed in crowded environments. Based on previous work on surface asperities [43], we hypothesized that modifying phage shape with molecules that bind to the phage filament surface could be a promising approach. We developed a minimal model and performed simulations with a mixture of phages, depletant biopolymers, and small Pf4 binders (S1 Fig and Methods). In our simulations, Pf4 phages are modeled as hard rods (via a linear array of hard discs), depletant biopolymers are modeled as hard discs, and Pf4 binders are modeled as hard discs with a single (fixed) circular patch that is attracted to cohesive patches on the Pf4 rod (S1A Fig). In the absence of binders, phage-phage proximity increases as depletant concentration increases (S1B Fig), indicating that our simulations capture the depletion attraction between phages in crowded environments, consistent with published work on several filamentous phages [25,31,32].

By adding Pf4 binders to the simulations, we were able to suppress this depletion attraction: in the presence of the binders, phage-phage proximity is reduced in comparison to simulations without binders, across a range of depletant concentrations (S1C Fig). Closer inspection of the rods in our simulations showed patchy and random coating of the binder along the sides of the phages (S1C Fig). We inferred that this roughened surface reduces the overlap excluded volume between the phages, ultimately leading to a reduction in the entropic gains of the excluded depletants away from the phage surface. Our simulations suggest that for the suppression of depletion, the binder must be large enough in comparison to the depletant particles (Figs 1A and S1D), in agreement with previous works on surface asperities [43–48]. In practice, binders may also modify the depletion interaction between phages via charge-based interactions with depletants (see Methods); however, we did not include charge in our simulations, because self-association of phages is generally observed to occur in charge-screened environments that suppress charge-based phage-phage repulsion [21,31]. Overall, our simulations suggest that a binder could suppress the depletion attraction between phages by increasing the average roughness of phage surfaces (Fig 1B), providing a potential mechanism to inhibit phage liquid crystalline droplets (Methods). We therefore predicted that a sub-stoichiometric Pf4 binder might be able to efficiently prevent Pf4 droplet formation, and thereby subvert the antibiotic tolerance conferred by the P. aeruginosa biofilm matrix.

Development of single-domain antibodies against Pf4

Since single-domain antibodies (nanobodies) have already shown promise in disrupting P. aeruginosa biofilms [40], we followed a similar approach to generate binders against Pf4. To this end, Pf4 phage was purified from P. aeruginosa PAO1 biofilms and used as an antigen to inoculate alpacas for nanobody generation. Nanobodies binding Pf4 were enriched by panning against biotinylated Pf4 immobilized on beads leading to 13 positive clones. To screen for Pf4 binders, nanobodies were purified, incubated with Pf4 filaments and subjected to a co-sedimentation assay. All nanobodies co-sedimented with Pf4 filaments to varying extents confirming binding (Figs 1C and S2A). Based on co-sedimentation behavior and sequence diversity, two nanobodies (Nb43 and Nb-D11) were carried forward for further characterization and development. Surface plasmon resonance (SPR) analysis confirmed that both Nb43 and Nb-D11 bind to Pf4, but with markedly different affinities (K_d = 69 ± 16 nM for Nb43 and 650 ± 60 nM for Nb-D11, Figs 1D and S2B–S2E). Differential scanning fluorimetry analysis indicated that both nanobodies were stable (Tm = 46.8 °C for Nb43 and 44.7 °C for Nb-D11), which was confirmed by SDS-PAGE analysis of nanobody incubated at both room temperature and 37 °C (S3 Fig). Next, we visualized Nb43 and Nb-D11 binding to Pf4 filaments using electron cryomicroscopy (cryo-EM). Pf4 mixed with Nb43 showed Pf4 filaments with a rugged surface, with speckled patchy nanobody decoration along the length of the filament as opposed to naked filaments seen in micrographs of Pf4 alone (Fig 1E, 1F). Two-dimensional (2D) class averages of Pf4 with Nb43 showed extra globular densities decorating the surface of the Pf4 filament, absent in class averages from a sample without nanobodies (Figs 1E, 1F, and S4A). Since the major coat protein of Pf4, CoaB, is the sole protein arranged along the entire outermost length of the phage, our analysis strongly suggests that Nb43 binds to CoaB. The Pf4 CoaB amino acid sequence is conserved in Pf1 and Pf6 phages and shows high sequence identity (87%) to the Pf5 major coat protein [49]. Furthermore, the amino acid sequence of all these CoaB proteins is identical in the N-termini, which form the outer surface of the phage that is recognized by Nb43, suggesting that Nb43 will bind to Pf phages from other P. aeruginosa strains in the same manner (S4H Fig). Interestingly, increasing Nb43 concentration resulted in Nb43 clustering on Pf4 filaments, leading to heavy decoration (Fig 1G, 1H). Using the previously estimated helical symmetry of Pf4 as a starting point [31], we performed three-dimensional (3D) reconstruction of Pf4 in complex with Nb43 (Figs 1I, 1J and S4A, S4B). Although we could only achieve a modest resolution even with heavily decorated filaments, it allowed us to model the spatial distribution of Nb43 along the Pf4 filament. This modeling from our structural data, together with a difference map between the Nb43+ Pf4 density and Pf4 without Nb43 (Pf4 alone) density, confirmed that the extra density, not accounted for by the Pf4 phage itself, was positioned outside the phage diameter directly on the surface of the phage, seen along the length of the phage filament. This analysis confirmed that Nb43 binds directly to CoaB, because the sole protein present along the length of the phage is CoaB (Figs 1I, 1J and S4D–S4G). In contrast to Nb43, Pf4 mixed with Nb-D11 showed no apparent nanobody decoration along Pf4 filaments in raw micrographs or in 2D class averages, suggesting either an alternative binding site for Nb-D11, or poorer binding due to lower affinity (S2D, S2E and S4C Figs).

Nanobody (Pf4) binders disrupt Pf4 liquid crystalline droplet formation

Pf4 phage liquid crystalline droplets that are abundant in the biofilm EPS matrix [21] can be recapitulated in vitro by mixing purified Pf4 with biopolymers, such as alginate, a polysaccharide which is also abundant in the P. aeruginosa biofilm matrix [31]. We used Alexa-488 (A488)-labeled Pf4, which allowed us to visualize the liquid crystalline droplets using fluorescence microscopy, enabling us to test the efficacy of Nb43 and Nb-D11 in preventing liquid crystalline droplet formation. We found that Nb43 was a potent inhibitor of liquid crystalline droplet formation in a dose-dependent manner, with spindle-shaped droplets almost entirely abolished at 1 μM Nb43 concentrations (Fig 2A–2C). The treated specimen showed mainly the formation of amorphous aggregates that did not resemble tactoids, suggesting that the ordered formation of spindle-shaped tactoids is impeded by Nb43, but formation of unordered aggregates still persists (Fig 2B). Nb-D11 also had an inhibitory effect; however, a 10-fold higher concentration (10 μM) was required for a comparable decrease in droplet formation (S5A–S5C Fig). Next, we assayed the ability of nanobodies to disrupt preformed droplets by pre-mixing Pf4 phage and alginate, allowing liquid crystalline droplets to form, before the addition of nanobodies. Both Nb43 and Nb-D11 could disrupt preformed liquid crystals effectively at similar concentrations to that required for prevention (1 μM for Nb43 and 10 μM for Nb-D11) (Figs 2D–2F and S5D–S5F). This inhibitory effect was specific against Pf4 droplets as both Nb43 and Nb-D11 did not prevent the formation of fd liquid crystalline droplets, a related filamentous phage from the Inovirus family that infects Escherichia coli [32] (S6A–S6F Fig). Furthermore, incubation of Pf4 droplets with an off-target nanobody, targeting the P. aeruginosa CdrA adhesin [40], did not affect Pf4 droplet formation, showing that inhibition requires Pf4 binding, and is not due to protein crowding effects (S6G–S6I Fig). Both Nb43 and Nb-D11 could also disrupt Pf4 liquid crystalline droplet formation when the depletant biopolymer was changed from alginate for dextran, a biopolymer with a different charge density [50], further suggesting that nanobodies act through binding to Pf4 filaments without requiring a charge-based interaction with depletant molecules (S7A–S7I Fig, Methods). Finally, we tested the ability of Nb43 to disrupt liquid crystalline droplets in the presence of synthetic cystic fibrosis sputum media (SCFM2), a medium containing mucin, DNA, and lipids providing a nutritional and physical environment similar to sputum [51]. Nb43 at 1 µM concentration could effectively disrupt Pf4 liquid crystalline droplets in this physiologically relevant medium, the same concentration as in other media (S7J–S7L Fig).

Next, using Alexa 568 (A568)-labeled Nb43, we probed the localization of Nb43 in Pf4 tactoids at low (noninhibitory) nanobody concentrations. Nb43 was homogeneously distributed throughout the liquid crystalline droplet (S8A–S8D Fig), correlating with the binding seen along single Pf4 filaments by cryo-EM (Fig 1). cryo-ET revealed the ultrastructure of Nb43 incubated Pf4 droplet samples, showing that Pf4 filaments were still associated with each other at 0.1 μM Nb43 even though Nb43 decoration was visible on Pf4 filaments (Fig 2G and S1 Movie). This showed that Nb43 can penetrate inside the liquid crystalline droplets and bind to phages. In contrast, at 1 μM Nb43 liquid crystalline droplet formation was prevented and Pf4 filaments that were heavily decorated with clustered nanobody were apparent (Fig 2H and S2 Movie). Parallel experiments with the other nanobody (A568-labeled Nb-D11) showed a distinct localization of Nb-D11, with Nb-D11 accumulating at the tips and edges of the spindle-shaped liquid crystalline droplets (S8E–S8H Fig). In line with these observations, cryo-ET showed droplets with a nearly normal appearance with disrupted ends (S8I Fig), confirming that Nb43 is more potent than Nb-D11, because it acts directly by binding along the phage at high affinity.

Nanobody binders disrupt Pf4 liquid crystalline droplet-mediated antibiotic tolerance by removing antibiotic diffusion block

We next studied the effect of Nb43 on Pf4 liquid crystalline droplet-mediated antibiotic tolerance. We have previously shown addition of Pf4 liquid crystalline droplets (i.e., Pf4 with alginate) was sufficient to encapsulate and protect P. aeruginosa cells from antibiotic-killing in vitro [31]. As nanobody binders could prevent and disrupt liquid crystalline droplets, we hypothesized that these nanobody binders could also affect antibiotic tolerance, as Pf4 liquid crystalline droplets were previously found to be protective against antibiotic treatment [21,31]. We tested this hypothesis by growing P. aeruginosa PAO1 ΔPA0728 (lacking the Pf4 integrase and unable to produce Pf4 phage [21]) in the presence of Pf4 liquid crystalline droplets, nanobody, and antibiotic. We confirmed that Pf4 droplets (Pf4 plus alginate) protect P. aeruginosa against aminoglycoside antibiotics in a bacterial survival assay, whereas Pf4 alone does not have any protective effect (Fig 3A–3B). Addition of 1 μM Nb43 abolished this protective effect (P_value < 0.05) and reduced P. aeruginosa survival to the level seen in the condition with no Pf4, but alginate alone (Fig 3A, 3B). Nb-D11 had a similar effect, but only at a 10-fold higher concentration (S9A, S9B Fig). In control experiments, both Nb43 and Nb-D11 had no effect on bacterial growth in isolation (Figs 3A, 3B and S9A, S9B). Given its higher affinity and potency against Pf4 liquid crystalline droplets, together with a defined site of binding on CoaB (Fig 1) that mirrors our biophysical modeling (Figs 1 and S1), and with a stronger effect in abolishing antibiotic tolerance, we went ahead with Nb43 for further experimentation.

Previously, we have shown that encapsulation of cells by Pf4 liquid crystalline droplets leads to a diffusion block that prevents antibiotic from reaching bacterial cells at an effective concentration [32]. We tested whether treatment with Nb43 relieves this diffusion block: fluorescently-labeled antibiotic (Texas Red-Gentamicin – GTTR) was added to P. aeruginosa cells in the presence or absence of Pf4 droplets and Nb43. Antibiotic uptake was monitored using the GTTR fluorescent signal, which slowly accumulated in the cells. In control cells lacking Pf4 liquid crystalline droplets, significant antibiotic uptake into cells could be detected (Fig 3C, 3D), which was diminished in the presence of Pf4 droplets, particularly in cells encapsulated by Pf4 liquid crystalline droplets (Fig 3E, 3F). Addition of 1 μM Nb43 restored antibiotic uptake to levels comparable to the control with no liquid crystalline droplets (Fig 3G, 3H), with a concomitant reduction in the number of cells encapsulated by Pf4 liquid crystalline droplets (Fig 3C–3J).

Nanobody binders increase antibiotic susceptibility of P. aeruginosa biofilms

Given that Nb43 could affect Pf4 liquid crystalline droplet formation, antibiotic tolerance, and antibiotic diffusion in vitro, we next tested whether this nanobody could affect antibiotic susceptibility of wild-type P. aeruginosa PAO1 biofilms producing endogenous Pf4. To this end, we tested the efficacy of Nb43 in static biofilms, where biofilm growth was measured using crystal violet staining in the presence of tobramycin and varying Nb43 concentrations (Fig 4A, 4B). Nb43 was added either at the beginning of growth (pre-incubation) or after 24 hours growth (without pre-incubation), with tobramycin being added at the 24 hours time point in both cases. Under pre-incubation conditions, 1 μM Nb43 significantly reduced biofilm growth by 50% compared to an antibiotic-only control at 1 μM Nb43 (P_value < 0.01). Without pre-incubation, a higher concentration of 2.5 μM Nb43 was required to obtain a similar reduction in biofilm growth as compared to the antibiotic treatment alone (P_value < 0.01) (Fig 4A, 4B). Next, we tested the ability of Nb43 to disrupt biofilms grown in the physiologically relevant synthetic sputum medium (SCFM2), with genomically characterized clinical isolates [52] where the Pf4 coat protein, CoaB, was present or absent in the genome. Under pre-incubation conditions, 1 μM Nb43 significantly reduced biofilm growth for the Pf4 containing clinical strain grown in SCFM2 (P_value < 0.01) and without pre-incubation a higher concentration of 2.5 μM was required (P_value < 0.01) (S10A, S10B Fig), mirroring the results obtained for PAO1 grown in LB medium (Fig 4A, 4B). Nb43 had no effect on the biofilm growth of the clinical isolate that lacked Pf4 at any concentration. Interestingly, the biofilms from the Pf4 lacking strain did show some tolerance to antibiotic, suggesting that another Pf4-independent mechanism of tolerance may be operating (S10 Fig).

To explore the effect of Nb43 on biofilms further, we used a custom microfluidics flow system to cultivate P. aeruginosa biofilms under flow, which we have used previously to assay antibiotic treatment of P. aeruginosa biofilms [40]. First, utilizing a P. aeruginosa strain expressing a fluorescent protein (mTFP – blue), we monitored Nb43 effect on biofilms in the absence of antibiotic. Nb43 treatment led to destabilization of biofilm growth and morphology even in the absence of antibiotic, with treated biofilms characterized by significantly less biofilm growth (P_value < 0.05) and a patchier cell distribution (P_value < 0.001) than biofilms grown in the absence of Nb43 (Fig 4C–4H). Next, we monitored the effect of Nb43 treatment on the antibiotic susceptibility of P. aeruginosa biofilms under flow. We used a sublethal tobramycin concentration for our experiments and monitored biofilm biomass and antibiotic susceptibility using P. aeruginosa cells constitutively expressing the Katushka2S fluorescent protein (far-red), together with BactoView dead staining (green) to identify dead cells. Biofilms treated with Nb43 and tobramycin showed significantly higher (P_value < 0.0001) cell death than biofilms treated with tobramycin alone or with no treatment (Fig 4I–4K), demonstrating that Nb43 treatment reduces antibiotic tolerance of P. aeruginosa biofilms. These results, taken together, confirm the efficacy of Nb43 in destabilizing P. aeruginosa biofilm morphology and increasing antibiotic susceptibility under various conditions (Fig 4), in line with our original theoretical modeling prediction that a phage binder could reduce phage self-association into liquid crystalline droplets (Figs 1 and S1).

Fig 5. Schematic representation of nanobody action in abolishing antibiotic tolerance of P. aeruginosa biofilms.

In untreated biofilms (left), cells show increased antibiotic tolerance due to Pf4 liquid crystalline droplets formed by depletion attraction in the biofilm EPS matrix, where encapsulated cells are protected by an antibiotic diffusion block. In nanobody treated biofilms, patchy binding of nanobody to Pf4 filaments reduces depletion attraction between the filaments preventing liquid crystalline droplet formation and encapsulation of cells, leading to increased antibiotic susceptibility of bacteria within the biofilm.

https://doi.org/10.1371/journal.pbio.3003834.g005

Biophysical methods: modeling of binder inhibition of depletion interaction

To model a system of phages, depletants, and binder particles, we used coarse-grained molecular dynamics simulations. We simulated a two-dimensional system consisting of N_r phages (with either N_r = 65 or N_r = 50 depending on the simulation) of length a = 80 nm and width b = 6.0 nm. We note that the chosen phage numbers and sizes are not precisely the same as that in the experiments and are typically smaller to minimize the significant increase in computational complexity; however, through such choices we are able to capture comparable phage densities and the role of their (long) aspect ratios (a/b>> 1). To represent each phage we used a rigid linear array of N_m = 17 overlapping discs of diameter b to ensure a sufficiently smooth surface (S1A Fig). Simulations also included N_d depletant discs of diameter , and N_b binder discs of diameter . The sizes of the depletant and binder particles were so chosen to efficiently explore, computationally, the role of the depletant interaction given the phage dimensions. To mediate the interaction between phages and binders, we included two smaller discs, or patches, of diameter fixed just beneath the surface of each disc on the phages, and one patch of diameter fixed on the circumference of each binder (S1A Fig). In the simulations, all N = 3N_rN_m + 2N_b + N_d particles were permitted to move in a box of fixed area L² with periodic boundary conditions at a fixed N and temperature T = 297 K (canonical ensemble).

In our model, the nonpatch particles experience excluded volume (steric) interactions, with the only attractive interaction occurring between the patch particles on each disc of the rod and the single patch on the binders—no charge-based interactions were included in the simulations. Any interaction between the patches on the rod is zero by construction. The total (sum over all pairs of particles) potential energy V accounting for these interactions is given as:

Equation 1

where is the set of all positions, is the distance between two particles i and j, = 10 k_BT is the excluded volume interaction strength, = 5 k_BT is the rod-patch to binder-patch cohesion strength, = 3  = 1.5 nm is the cohesion range, denotes the particle type of i (1 = rod, 2 = depletant, 3 = binder, 4 = binder-patch, and 5 = rod-patch), factors of is a way to only impose excluded volume interactions to types 1–3 and cohesive interactions to types 4–5, is the Kronecker delta function, H(…) is the Heaviside function, and is the Lorentz-Berthelot rule with being the diameter of a particle of type t(i). imposes excluded volume and is given by the Weeks-Chandler-Andersen potential as:

Equation 2

The cohesive interaction, , is determined by a truncated and shifted Lennard-Jones potential given as:

Equation 3

where is the force of the Lennard-Jones potential.

To obtain dynamical trajectories of the above model, we used the following protocol. We time integrate overdamped Langevin dynamics in the LAMMPS (Aug 2024) package [58], with a timestep of ns. We impose the rigidity of the rods through a constraint term in the dynamics, as specified by the fix rigid command in LAMMPS. Initially, the system is prepared with the rods uniformly spread across the plane in grid-like fashion, and then a small initial simulation is conducted dragging depletant and binder particles into the simulation box of area L² = 600 x 600 nm². Before any recording of the observables, the system is equilibrated for 10⁷ timesteps and production runs last timesteps. Equilibration is checked through inspection of the total energy and the averaged orientation of the rods.

Biophysical modeling considerations: mechanism of binder inhibition of Pf4 liquid crystalline droplet formation

In previous work, the phage Pf4 has been found to form protective liquid crystalline droplets around bacterial cells in environments that are charge-screened and crowded; such environments suppress charge-based phage-phage repulsion and generate a depletion attraction between phages [21,31]. Therefore, in this study, we applied our model to explore a method to suppress a nonspecific interaction induced by crowding agents, rather than adopting the usual route of blocking an active site of an enzyme or blocking a protein:protein interaction. As described in the main text, our biophysical modeling results (Figs 1 and S1) suggested that a sub-stoichiometric binder could disrupt Pf4 phage liquid crystalline droplets through surface roughening, which dampens the depletion attraction. Our model predicts that such a mechanism would require the binder being large enough in comparison to the depletant (Fig 1A). In our experiments, we found that Nb43, a nanobody that binds to Pf4 and decorates its surface (Fig 1), suppresses the formation of liquid crystalline droplets (Fig 2), in line with our theoretical prediction of reduced Pf4 self-association because of surface roughening. At higher concentrations of Nb43, we observed clustering of the binder on Pf4 filaments, leading to heavy decoration (Fig 1E, 1F), which may contribute to the effect of the surface roughening by increasing its effective size.

In addition to the surface roughening mechanism, the nanobody Nb43 may also reduce the effective size of the depletant, alginate, in the phage-phage depletion interaction. This would involve the phage-bound nanobodies either screening charge-based repulsion between phages and depletants (both alginate and phages are negatively charged) or attracting depletants (the nanobodies are overall positively charged): our simulations do not account for these charge-based interactions. However, because our experiments were performed in a charge-screened environment, it is unlikely that this is a major factor in the nanobody’s mechanism. Furthermore, the nanobody’s action was found to be replicated in experiments with dextran as a depletant, which carries less overall negative charge (S7A–S7I Fig). Overall, the combined modeling and experimental results suggest that suppression of depletion via surface roughening, as predicted by our model, is the main mechanism by which the nanobody Nb43, a Pf4 binder, disrupts Pf4 liquid crystalline droplet formation.

Bacterial strains

P. aeruginosa strains PAO1 or PAO1 ΔPA0728 (a kind gift from Prof. Patrick Secor, Montana State University) or PA14:mTFP (modified to express teal fluorescent protein (mTFP) under the PA1/04/03 promoter at the att7 site) or the PA14 att7::Katushka2S were utilized as indicated. E. coli WK6 was used for periplasmic expression of nanobodies. Shaking cultures of bacteria were grown at 37 °C in Luria-Bertani (LB) medium with agitation at 180 rpm (revolutions per minute) unless otherwise specified.

Bacterial media

The following bacterial media were used for experiments. LB: Composed of 5 g/L yeast extract, 10 g/L tryptone, and 5 g/L NaCl, dissolved in milliQ water. Terrific broth (TB): 12 g/L tryptone, 24 g/L yeast extract, 0.4% (v/v) glycerol, 0.017 M potassium dihydrogen phosphate and 0.072 M potassium phosphate dibasic. KA biofilm medium [59]: Composed of 6 g/L Tris-HCl (adjusted to pH 7.4), 4 g/L L-arginine HCl, 174 mg/L dipotassium sulfate, and 72 mg/L magnesium sulfate in water. SCFM2 was prepared as described previously [51].

Pf4 phage amplification

Pf4 bacteriophage was produced as described previously [31]. Briefly, Pf4 was amplified from stock Pf4 isolated from PAO1 biofilms [31] by incubating 1 x 10³ pfu (plaque forming units)/ml of Pf4 stock with 1 ml of PAO1 culture at 0.5 optical density measured at 600 nm wavelength of light (OD₆₀₀₎ for 15 min before mixing with hand-hot 0.8% (w/v) agar. The mixture was plated onto 10 cm² LB-agar plates and incubated overnight at 37 °C. Each plate was covered with 5 ml phosphate buffered saline (PBS) and incubated for 6 hours at room temperature (RT) before the PBS was collected and centrifuged (12,000 g, 30 min, 4°C) to remove debris. Pf4 was precipitated from the supernatant by polyethylene glycol (PEG) precipitation. The supernatant solution was adjusted to 0.5 M NaCl and Pf4 precipitated by addition of 10% (w/v) PEG 6,000 (Sigma) followed by overnight incubation at 4 °C. Precipitated Pf4 phage particles were pelleted by centrifugation (12,000 g, 30 min, 4 °C), resuspended in PBS and dialyzed overnight against PBS using 10 kDa MWCO (molecular weight cutoff) snakeskin dialysis membranes (Thermo Fisher Scientific). Yield of the Pf4 bacteriophage preparation was estimated using Nanodrop (Thermo Fisher Scientific)

Nanobody generation, expression, and purification

Nb-D11 was generated by the Nanobody Discovery Platform, Rosalind Franklin Institute, UK and Nb32–45 were generated by the VIB Nanobody Service Facility, Belgium, by methods previously published [60,61]. Plasmids encoding His₆-tagged-nanobodies were transformed (separately) by heat shock into E. coli WK6 cells. TB supplemented with 100 µg/ml ampicillin was inoculated with transformed E. coli WK6 cells and incubated at 37 °C with agitation at 180 rpm. Upon reaching an OD₆₀₀ of between 0.6 and 0.8, cultures were induced with 1 mM isopropyl β-D-thiogalactoside (IPTG) and incubated overnight at 23 °C with agitation at 180 rpm. Cells were pelleted by centrifugation (5,000 g, 10 min, 4 °C) and resuspended in 15 ml ice-cold Tris/Ethylenediaminetetraacetic acid (EDTA)/Sucrose (TES buffer – 200 mM Tris, 650 μM EDTA and 500 mM sucrose) per liter of culture and incubated for 1 hour with agitation on an orbital rocker. To disrupt the periplasm, 30 ml ice-cold TES/4 (1 part TES buffer and 3 parts milliQ water) was added per liter of culture and agitated on an orbital rocker for 45 min at 4 °C. After incubation, debris was pelleted by centrifugation (10,000 g, 30 min, 4 °C), the supernatant recovered and subjected to Nickel-NTA affinity chromatography with a 5 ml His-Trap high performance (HP) column (Cytiva) followed by size exclusion chromatography with a Superdex S75 column (Cytiva) in size exclusion buffer (50 mM Tris (pH 7.4), 150 mM NaCl, 0.5 mM TCEP). Nanobody containing fractions were concentrated and the protein concentration was estimated using a nanodrop (Thermo Fisher Scientific). Nanobodies were aliquoted, snap-frozen in liquid nitrogen and stored at −80 °C ahead of use.

Nanobody co-sedimentation with Pf4 filaments

Nanobody was centrifuged (100,000 g, 1 hour, 4 °C) to remove any aggregates formed during storage and the supernatant used in the co-sedimentation assay after further protein concentration determination by nanodrop (Thermo Fisher Scientific). Each nanobody (1 µg) was incubated either alone or with an equal amount (w/w) of purified Pf4 in a 100 μl reaction volume for 1 hour at room temperature before centrifugation (100,000 g, 1 hour, 4 °C). The supernatant was carefully removed from the pellet, which was resuspended in the same volume of PBS as the supernatant. Equivalent volumes of the supernatant (S) and pellet fractions were loaded onto 4%–12% tris-glycine Novex gels (Thermo Fisher Scientific) and proteins visualized by Coomassie Blue staining.

SPR

SPR was performed using a Biacore T200 using C1-sensor chip (Cytiva). Both reference control and analyte channels were equilibrated HBS-N buffer (Cytiva). Pf4 was immobilized onto the chip surface via amide coupling using the supplied kit (Cytiva) to reach a RU value of 400 RU. SPR runs were performed with analytes injected for 120 s followed by a 300 s dissociation in a 1:2 dilution series with initial concentrations of 10 µM. After reference and buffer signal correction, sensorgram data were fitted using Prism Graphpad (GraphPad Software Inc). The equilibrium response (R_eq) data were fitted to a single site interaction model with linear nonspecific binding to determine K_d:

Equation 4

where C is the analyte concentration and R_max is the maximum response at saturation, N is the nonspecific binding and B is the background resonance.

Differential Scanning Fluorimetry

Thermal denaturation was followed using intrinsic protein fluorescence measured with the NanoTemper Prometheus NT48 instrument (Nanotemper Technologies, Germany). Nanobody samples at 2 µM concentration in 50 mM Tris pH 7.5, 150 mM NaCl, and 0.5 mM TCEP buffer were loaded into standard capillaries and heated at 2 °C/min from 20 to 95 °C. The first derivative of the fluorescence emission ratio 350/330 nm was analyzed using the PR.ThermControl v2.3.1 (NanoTemper), to define the T_m.

Cryo-EM grid preparation

Samples for cryo-EM were prepared as described previously [31], by pipetting 2.5 μl of the sample onto freshly glow- discharged Quantifoil grids (Cu/Rh R2/2, 200 mesh or Au R1.2/1.3, 300 mesh for cryo-EM, Cu/Rh R3.5/1, 200 mesh for cryo-ET) and plunge-frozen into liquid ethane in a Vitrobot Mark IV (Thermo Fisher Scientific). Plunge-frozen grids were transferred to liquid nitrogen and stored until imaging.

Cryo-EM and cryo-ET data collection

Cryo-EM data for screening specimens were collected either using a Titan Krios or Glacios microscope (Thermo Fisher Scientific) operated at 300 kV or 200 kV, respectively. High-throughput data were collected using the EPU software on a Titan Krios microscope fitted with a Selectris energy filter (slit width 10 eV) and a Falcon 4i direct electron detector operating in counting mode at an unbinned, calibrated pixel size of 0.96 Å. A combined total dose of ~50 e^-/Ų was applied with each exposure lasting 4 s and 40 frames were recorded per movie. In total, 3,601 movies were collected between −1 and −2.5 μm defoci. Tilt series data for cryo-ET were collected on a Titan Krios using the BioQuantum energy filter (Gatan) and K3 direct electron detector (Gatan) with the SerialEM software [62]. Tilt series were collected in a dose-symmetric scheme starting from 0° between ± 60° with 1° tilt increment between −6 and −8 μm defoci, with a combined dose of ~163 e-/Ų applied over the entire series. Tilt series were collected at an unbinned, calibrated pixel size of 2.13 Å. Acquired tomographic data were pre-processed in RELION5 [63], and tomograms were reconstructed using AreTomo [64] and denoised using cryoCARE [65,66].

Cryo-EM data processing

Cryo-EM image processing of 2D image data of nanobody-decorated phage filaments was performed in cryoSPARC [67]. After patch motion correction and CTF estimation, filament positions were determined within micrographs using the cryoSPARC filament picker, extracting particles every 1.2 nm along the filament axis, resulting in one unique nanobody binder in each box according to estimates based on previous 2D class averages. 2D class averages were then created in cryoSPARC. A wide range of symmetries for helical refinement was tested, but did not yield densities with resolved Nb43-bound CoaB subunits. 3D reconstructions of Nb43-bound Pf4 were ultimately generated using helical refinement in cryoSPARC with no symmetry applied. For visualization of the densities in the cryo-EM map of Nb43-bound Pf4 that do not correspond to the Pf4 filament itself, a model of Pf4 was fitted into the density, and a cylindrical mask of 70 Å was centered onto the atomic model of the Pf4 filament. A density comprising only non-Pf4 moieties was created by subtracting this cylindrical mask from the Nb43-bound Pf4 cryo-EM density using relion_image_handler, resulting in a map containing only external densities, which is displayed in orange.

Fluorescent labeling of Pf4 and nanobody

Purified Pf4 or nanobody was dialyzed into PBS using 10 kDa MWCO snakeskin dialysis membrane (Thermo Fisher Scientific). To adjust the pH of the sample, 100 μl of 1 M sodium carbonate buffer pH 9.2 was added to 900 μl Pf4 phage (5 mg/ml) or 900 μl nanobody (10 mg/ml). The pH-adjusted Pf4 sample was incubated with 100 µg Alexa fluor 488-NHS fluorescent dye (Thermo Fisher Scientific) for 1 hour at RT with end-over-end agitation. The pH-adjusted nanobody samples were incubated with 100 µg Alexa fluor 568-NHS fluorescent dye (Thermo Fisher Scientific). Both A488-labeled phage and A568-labeled nanobody were isolated from free dye by passing over two PD10 desalting columns (Cytiva). Protein concentration was estimated by nanodrop (Thermo Fisher Scientific).

Fluorescence microscopy

Texas Red-gentamicin (GTTR) diffusion into bacterial cells

PAO1 ΔPA0728 cells were grown to an OD₆₀₀ of 0.5 and incubated with A488-labeled phage (final concentration 1 mg/ml) and sodium alginate (final concentration 4 mg/ml) and 1 μM Nb43 for 1 hour. Texas Red-gentamicin (AAT Bioquest) was added to a final concentration of 1 µM and incubated for a further 4 hours. 5 µl of the sample was pipetted onto 1.5% (w/v) agar pads constructed using 15 x 16 mm Gene Frames (ThermoFisher) with a coverslip applied before subsequent imaging using a Zeiss Axioimager M2 microscope (Carl Zeiss) in both brightfield and fluorescent mode. Presented images are background subtracted and figure panels were prepared using Fiji [69]. Images were quantified by semi-automated segmentation of bacterial cells and associated Pf4 liquid crystalline droplet (see below), and fluorescent intensity in the Texas Red channel was measured at the coordinates of segmented bacteria. Graphs were plotted using Prism Graphpad (GraphPad Software Inc).

Semi-automated segmentation of images with bacterial cells encapsulated by Pf4 droplets

Bacterial cells were selected from the brightfield channel images to locate the positions of cells. Bacterial cell shapes were found using the activecontour algorithm in MATLAB [70]. The regions of identified bacteria were dilated and used as seed inputs for the segmentation of the Pf4 liquid crystalline droplets in the fluorescence channel using the activecontour algorithm. The segmented bacterial cells from the brightfield and the Pf4 liquid crystalline droplets from the fluorescence channel were then used to calculate encapsulation. Fluorescent intensity in the Texas Red channel was measured at the coordinates of segmented bacteria.

Nanobody disruption of Pf4 liquid crystalline droplet-mediated antibiotic protection in vitro

An overnight culture of PAO1 ΔPA0728 was grown in LB media at 37 °C, diluted 1 in 100 into fresh LB medium and grown at 37 °C to an OD₆₀₀ of 0.5. 100 µl of the resulting culture was added to a 96-well plate and incubated further for 30 min at 37 °C. 100 µl of Pf4 and/or alginate, and nanobody were added to the culture such that final concentrations of components were: sodium alginate (Scientific Laboratory Supplies) (4 mg/ml), Pf4 (1 mg/ml) and nanobody at the indicated concentration. Additionally, tobramycin (10 µg/ml) (Sigma) or gentamicin (10 µg/ml) (Sigma) was added as indicated and cultures were grown further for 3 hours. A 10 µl sample for each assay condition was serially diluted 10-fold and 100 µl of the dilutions plated onto LB-agar plates. Plates were inverted and incubated overnight at 37 °C and colonies forming units (cfu) were enumerated. Experiments were performed in triplicate. Mean cfu/ml with standard deviation were calculated and plotted using the Prism GraphPad software (GraphPad Software Inc).

Antibiotic susceptibility of static P. aeruginosa biofilms

Static biofilm assays were adapted from a previously described protocol [71]. An overnight culture of PAO1 or clinical isolates containing or lacking Pf4 major coat protein, CoaB, was grown in either LB media or SCFM2 at 37 °C. The culture was diluted 1 in 100 into fresh LB media or SCFM2 and 100 µl pipetted into each well of a flat-bottomed 96-well plate. In the case of pre-incubation experiments, nanobody at the indicated concentration was added to the well. Plates were incubated overnight at 37 °C before addition of 1 µg/ml tobramycin and incubation for a further 8 hours. For experiments without a pre-incubation, Nb43 was added at the indicated concentration at the same time as tobramycin. Subsequently, planktonic cells were removed from the plate by inverting and shaking out the culture. The plate was washed three times by submerging in milliQ water and shaking out contents over paper towels. To visualize biofilms, 200 µl 0.1% crystal violet (Sigma) was added to each well and incubated at room temperature for 15 min. Plates were washed a further three times with MilliQ water as described previously and plates were inverted and left to dry overnight. To quantify biofilm formation, 200 µl 30% (v/v) acetic acid was added to each well and incubated for 15 min. Samples were transferred to a fresh 96-well plate and crystal violet signal was quantified by measuring absorbance at a wavelength of 550 nm in a Tecan Infinite M200 Pro plate reader. Experiments were repeated four times. Mean absorbance with standard deviation was calculated and plotted using Prism GraphPad software (GraphPad Software Inc).

Nanobody effect on P. aeruginosa biofilms grown under flow conditions

S1 Fig. Minimal biophysical modeling suggests that Pf4 binders can disrupt depletion attraction between Pf4 phages.

(A) Components of the coarse-grained molecular dynamics model (top): phages are modeled as hard rods of length a = 80.0 nm, and width b = 6.0 nm, the depletant particles are modeled as hard discs with a diameter s_d = 2.4 nm and the binders are modeled as hard discs of diameter s_b = 2.4 nm with a single (fixed) circular binding patch that is only attracted to the circular binding patches on the phage rod. Visual depiction of phage-phage alignment and localization, here termed proximity (quantified below). (B) Simulation snapshots of 65 phages with increasing numbers of depletant particles (100, 8,000, 16,000, 20,000) from left to right, phages show greater proximity for increased numbers of depletant particles. Scale bar is 100 nm. (C) The presence of binders significantly disrupts depletion attraction between rods. (i) Same as (B) but with a constant presence of 5,000 binders. Despite the increase in depletant numbers, phage proximity remains low. (ii) Increasing the number of binders (400, 625, 1,250, and 2,500) from left to right with 20,000 depletant particles. Even a small number of binders disrupts the depletion interaction between rods. (D) Quantification of phage-phage proximity, the number of aligned phage pairs whose centers are within an approximate phage width divided by twice the number of phages, as a function of depletant density. Lines show different ratios of the depletant diameter and phage width. In panel D, error bars on plots from averaging over multiple time points and simulations are too small to be visible. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s001

(TIF)

S2 Fig. Nanobody binding to Pf4 filaments.

(A) Coomassie-stained SDS-PAGE of the soluble (S) and pellet (P) fractions from Pf4 filament co-sedimentation assay with a panel of recombinant nanobodies. Nanobodies (1 μg) were incubated alone or in the presence of Pf4 (1 μg) as indicated prior to centrifugation at 100,000 g. Without Pf4, nanobodies remained predominantly in the soluble fraction, but nanobodies incubated with Pf4 co-sedimented with Pf4 in the pellet fraction after centrifugation, indicating binding to Pf4. Arrows indicate bands corresponding to nanobody (Nb) and the Pf4 major coat protein (CoaB). Molecular weight markers are shown on the left. Red asterisks denote nanobodies selected for further characterization (Nb43 and Nb-D11). (B–E) Pf4 was immobilized via amide coupling onto a C1-sensor chip and nanobodies were flowed over the chip. (B) Nb43 sensorgram and (C) response curve showing Nb43 has a low nanomolar affinity for Pf4 (K_d = 69 ± 16 nM), (D) Nb-D11 sensorgram and (E) response curve showing Nb-D11 has a high nanomolar affinity for Pf4 (K_d = 650 ± 60 nM). The data underlying this Figure can be found in S1 Data and S1 Raw Images.

https://doi.org/10.1371/journal.pbio.3003834.s002

(TIF)

S3 Fig. Nanobody stability.

(A) Differential scanning fluorimetry of Nb43. Melting curve for 1 µM Nb43 over a temperature range of 20–95 °C. Intrinsic fluorescence was measured as the ratio of fluorescence at 350 nm/ 330 nm. Nb43 has a Tm of 46.8 °C.(B) Time-course of Nb43 stability at room temperature (RT) and 37 °C. Nb43 was incubated at the indicated temperature for the indicated time before analysis by SDS-PAGE to assay for protein degradation. (C) Differential scanning fluorimetry of Nb-D11. Melting curve for 1 µM Nb-D11 over a temperature range of 20–95 °C. Intrinsic fluorescence was measured as the ratio of fluorescence at 350 nm/ 330 nm. Nb-D11 has a Tm of 44.7 °C. (D) Time-course of Nb-D11 stability at RT and 37 °C. Nb43 was incubated at the indicated temperature for the indicated time before analysis by SDS-PAGE to assay for protein degradation. The data underlying this Figure can be found in S1 Data and S1 Raw Images.

https://doi.org/10.1371/journal.pbio.3003834.s003

(TIF)

S4 Fig. Cryo-EM analysis of Pf4 filaments incubated with nanobodies.

(A) Gallery of 2D class averages illustrating heterogeneity of Nb43 occupancy on Pf4 filaments. (B) Resolution estimation of Pf4-Nb43 reconstruction (Fig 1I) by Fourier shell correlation (FSC) of independently aligned and averaged half-maps. The dashed line indicates 0.143 criterion. (C) Cryo-EM micrograph of Pf4 with Nb-D11, showing Pf4 filaments with no apparent decoration. The inset shows a representative 2D class average. (D) Schematic representation of Pf4 phage showing distribution of Pf4 structural proteins. (E) Cryo-EM density of Pf4 alone (EMDB-10593, low pass filtered to 15 Å) with model fitted in blue. (F) Schematic representation of Pf4 phage showing distribution of Pf4 structural proteins bound to Nb43. (G) Map of Pf4-Nb43 with density corresponding to the Pf4 filament coloured blue, and Nb43 density coloured orange. Nb43 density is seen decorating the surface of Pf4 phage showing direct binding to Pf4 CoaB, the sole protein present along the phage filament surface. (H) Clustal Omega multiple sequence alignment of CoaB proteins from different Pf phages. * = sequence identity, : = strong sequence homology, . = weak sequence homology. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s004

(TIF)

S5 Fig. Nb-D11 prevents and disrupts Pf4 liquid crystalline droplets.

(A and B) Representative light microscopy images of A488-labeled Pf4 liquid crystalline droplets formed in the presence of (A) no Nb-D11 or (B) 10 μM Nb-D11. (C) Bar chart showing the abundance of Pf4 liquid crystalline droplets per μm². Addition of 10 μM Nb-D11 results in a statistically significant reduction in liquid crystalline droplet formation (P_value < 0.01). (D and E) Representative light microscopy images of preformed liquid crystalline droplets treated with (D) no Nb-D11 or (E) 10 μM Nb-D11. (F) Bar chart showing the abundance of Pf4 liquid crystalline droplets per μm² after treatment. Addition of 10 μM Nb-D11 results in a statistically significant reduction in droplet abundance (P_value < 0.01). Error bars represent standard deviation. P-values were calculated using an unpaired t test. All images have been background subtracted. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s005

(TIF)

S6 Fig. Disruption of Pf4 liquid crystalline droplets requires specific nanobody binding to Pf4 filaments.

(A and B) Representative light microscopy images of A488-labeled fd liquid crystals formed in the presence of (A) no Nb43 or (B) 10 μM Nb43. (C) Bar chart showing the abundance of fd liquid crystals per μm². Addition of 10 μM Nb43 did not result in a statistically significant change in liquid crystal formation. (D and E) Representative light microscopy images of A488-labeled fd liquid crystals formed in the presence of (D) no Nb-D11 or (E) 10 μM Nb-D11. (F) Bar chart showing the abundance of fd liquid crystals per μm². The addition of 10 μM Nb-D11 did not result in a statistically significant change in liquid crystal formation. (G and H) Representative light microscopy images of A488-labeled Pf4 liquid crystalline droplets formed in the presence of (G) no nanobody or (H) 10 μM anti-CdrA nanobody. (I) Bar chart showing the abundance of Pf4 liquid crystalline droplets per μm². The addition of 10 μM anti-CdrA nanobody did not result in a statistically significant change in liquid crystalline droplet formation. All values are representative of 30 images from three independent replicates. Error bars represent standard deviation. P-values were calculated using an unpaired t test. All images have been background subtracted. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s006

(TIF)

S7 Fig. Nanobodies disrupt Pf4 liquid crystalline droplet formation with different media and depletants.

(A and B) Representative light microscopy images with comparable total liquid crystalline droplet area formed by A488- labeled Pf4 using (A) alginate and (B) dextran as the depletant. (C) Bar chart showing total liquid crystalline droplet area as assessed by light microscopy followed by image segmentation of droplets. (D ad E) Representative light microscopy images of A488-labeled Pf4 liquid crystalline droplets formed with dextran as the depletant in the presence of (D) no Nb43 or (E) 1 μM Nb43. (F) Bar chart showing the abundance of Pf4 liquid crystalline droplets per μm². Addition of 1 μM Nb43 results in a statistically significant reduction in liquid crystalline droplet formation (P_value < 0.01). (G and H) Representative light microscopy images of A488-labeled Pf4 liquid crystalline droplets formed with dextran as the depletant in the presence of (G) no Nb-D11 or (H) 10 μM Nb-D11. (I) Bar chart showing the abundance of Pf4 liquid crystalline droplets per μm². Addition of 10 μM Nb-D11 results in a statistically significant reduction in liquid crystalline droplet formation (P_value < 0.01). (J–L) Representative light microscopy images of A488-labeled Pf4 liquid crystalline droplets formed in synthetic sputum medium SCFM2 with alginate as the depletant in the presence of (J) no Nb43 or (K) 1 μM Nb43. (I) Bar chart showing the abundance of Pf4 liquid crystalline droplets per μm². Addition of 1 μM Nb43 results in a statistically significant reduction in liquid crystalline droplet formation (P_value < 0.001). The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s007

(TIF)

S8 Fig. Nanobody binder localization in Pf4 liquid crystalline droplets.

(A–C) Representative light microscopy images of A488-labeled Pf4 incubated with alginate and 0.1 µM A568-labeled Nb43, (A) A488-Pf4 liquid crystalline droplets, (B) A568-labeled Nb43, and (C) merge. Images have been background subtracted. (D) Fluorescent intensity line profile through cross-section of Pf4 liquid crystalline droplet showing homogeneous signal for Nb43 (red line) throughout liquid crystalline droplets (green line). (E–G) Representative light microscopy images of A488-labelled Pf4 incubated with alginate and 2.5 µM A568-labelled Nb-D11. Images have been background subtracted. (H) Fluorescent intensity line profile through cross-section of Pf4 liquid crystalline droplets showing signal for Nb-D11 (red line) is excluded to the tips and surface of the droplet (green line). (I) Cryo-ET slice of a Pf4 liquid crystalline droplet incubated with 2.5 µM Nb-D11. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s008

(TIF)

S9 Fig. Nb-D11 disrupts phage liquid crystalline droplet-mediated antibiotic tolerance in vitro at high concentrations.

(A and B) Bar graph shows colony-forming units (cfu) per ml (y axis), a measure of P. aeruginosa culture cell viability after (A) tobramycin and (B) gentamicin treatment in the presence of different reagents (x axis). For both antibiotics, 10 µM Nb-D11 significantly reduced antibiotic tolerance levels to that seen in the alginate alone condition (P_value < 0.05). Values shown are the mean of three independent experiments. Error bars represent standard deviation. P-values were calculated using an unpaired t test. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s009

(TIF)

S10 Fig. Nb43 action on biofilms of clinical isolates grown in synthetic sputum media.

(A and B) A P. aeruginosa clinical strain encoding Pf4 major coat protein, CoaB, was used to grow static biofilms in 96-well plates in SCFM2 media and treated with Nb43 either (A) at the inoculation stage (pre-incubation) or (B) after 24 hours growth (without pre-incubation) with the indicated concentrations of Nb43. At the 24 hours time point, 1 μg/ml tobramycin was added and cultures incubated for a further 8 hours, before plates were treated with crystal violet to assay for biofilm growth. Significantly less biofilm is present with (A) 1 μM Nb43 in pre-incubation conditions (P_value < 0.01) and (B) 2.5 μM Nb43 in the without pre-incubation condition (P_value < 0.01) compared to the control without Nb43. (C and D) A P. aeruginosa clinical strain lacking Pf4 major coat protein, CoaB, was used to grow static biofilms in 96-well plates in SCFM2 media and treated with Nb43 either (C) at the inoculation stage (pre-incubation) or (D) after 24 hours growth (without pre-incubation) with the indicated concentrations of Nb43. At the 24 hours time point, 1 μg/ml tobramycin was added and cultures incubated for a further 8 hours, before plates were treated with crystal violet to assay for biofilm growth. Graphs show absorbance at 550 nm (crystal violet signal indicating amount of biofilm) on the y axis and components added to the assay on the x axis. Addition of Nb43 had no effect on biofilms formed with the clinical strain lacking CoaB in both (C) pre-incubation and (D) without pre-incubation conditions as compared to the control without Nb43. In all cases, mean values from three replicates are plotted. Error bars represent standard deviation. P-values were calculated using an unpaired t test. The data underlying this Figure can be found in S1 Data.

https://doi.org/10.1371/journal.pbio.3003834.s010

(TIF)