Pushing beyond the Envelope: the Potential Roles of OprF in Pseudomonas aeruginosa Biofilm Formation and Pathogenicity

The ability of Pseudomonas aeruginosa to form biofilms, which are communities of cells encased in a self-produced extracellular matrix, protects the cells from antibiotics and the host immune response. While some biofilm matrix components, such as exopolysaccharides and extracellular DNA, are relatively well characterized, the extracellular matrix proteins remain understudied.

ABSTRACT

The ability of Pseudomonas aeruginosa to form biofilms, which are communities of cells encased in a self-produced extracellular matrix, protects the cells from antibiotics and the host immune response. While some biofilm matrix components, such as exopolysaccharides and extracellular DNA, are relatively well characterized, the extracellular matrix proteins remain understudied. Multiple proteomic analyses of the P. aeruginosa soluble biofilm matrix and outer membrane vesicles, which are a component of the matrix, have identified OprF as an abundant matrix protein. To date, the few reports on the effects of oprF mutations on biofilm formation are conflicting, and little is known about the potential role of OprF in the biofilm matrix. The majority of OprF studies focus on the protein as a cell-associated porin. As a component of the outer membrane, OprF assumes dual conformations and is involved in solute transport, as well as cell envelope integrity. Here, we review the current literature on OprF in P. aeruginosa, discussing how the structure and function of the cell-associated and matrix-associated protein may affect biofilm formation and pathogenesis in order to inform future research on this understudied matrix protein.

INTRODUCTION

Two million people contract nosocomial infections each year, extending hospitalization times and associated costs. Pseudomonas aeruginosa is one of the leading causes of such infections (1). This bacterium possesses high intrinsic and acquired antibiotic resistance, which complicates treatment and contributes to the 16% mortality rate associated with P. aeruginosa septicemia (2, 3). This versatile opportunistic human pathogen can infect patients undergoing burn wound treatment, organ transplantation, cancer therapy, prosthetic implantation, artificial ventilation, and catheterization (4). However, the notoriety of P. aeruginosa is due to its colonization of the respiratory tract of cystic fibrosis patients. This Gram-negative bacterium commonly colonizes the airways of young cystic fibrosis patients, and this infection often persists throughout their lives. Chronic P. aeruginosa infection dominates end-stage cystic fibrosis lung disease, resulting in respiratory failure and mortality (5).

While P. aeruginosa employs many strategies to evade antibiotics and host immune defenses, one of its most effective survival techniques is the formation of bacterial communities, called biofilms. Bacteria in biofilms are significantly more tolerant against bactericidal agents and host immune responses than their planktonic counterparts (4). Biofilms are encased in a protective matrix, which is composed of exopolysaccharides, proteins, lipid vesicles, and extracellular DNA (eDNA). Biofilm development progresses in five stages: (i) reversible attachment, in which motile (planktonic) cells attach to a surface using the flagellum; (ii) irreversible attachment, where cells become more firmly connected to the surface via the long axis of the cell; (iii) microcolony formation, in which cells aggregate and secrete matrix components; (iv) mature biofilm, characterized by macrocolony and fluid channel formation; and (v) dispersal (6). Distinct physiology and gene expression profiles exist between P. aeruginosa cells in each stage (7–9). Furthermore, biofilm development is influenced by quorum sensing and cyclic diguanylate (c-di-GMP) signaling (10), as well as environmental factors such as nutrient availability, oxygen concentration, and fluid shear (11).

Previous proteomic analyses of P. aeruginosa have detected the outer membrane (OM) porin OprF in the biofilm matrix and in outer membrane vesicles (OMVs) (12–14). Porins are integral OM proteins that form hydrophilic channels through which charged solutes can pass (15, 16). While OprF has been extensively studied as a component of the OM, the role of OprF in biofilm formation is relatively understudied. It is unknown whether the conformations, functions, and associations of the cell-associated porin occur in OprF within the matrix and how they impact the P. aeruginosa biofilm. We review here the current OprF literature and discuss the potential roles of the porin in the biofilm matrix, OMVs, the mammalian immune response, and P. aeruginosa virulence.

SLOW AND STEADY: CELL-ASSOCIATED STRUCTURE AND FUNCTION OF OprF

In 1979, Hancock et al. identified an abundant 35-kDa OM protein in P. aeruginosa, which was originally referred to as protein F and would later be renamed OprF (17). OprF is present in high abundance, with approximately 100,000 to 300,000 copies per P. aeruginosa cell (18). During liposome reconstitution, the suspected porin exhibited unusually slow permeability to hydrophilic solutes (17). The permeability rate of OprF was 40-fold lower than that of the known Escherichia coli porin OmpF (19). While the porin is slow, OprF allows passage of large solutes up to 3 kDa in size through the OM (20, 21). For comparison, the general porins known at the time for E. coli, which harbors a far more permeable OM, are permissible to solutes with molecular weights of 500 Da or less (15).

The similarities between OprF and the E. coli OM porin OmpA were later uncovered, when Duchêne et al. reported in 1988 that a 30-amino-acid region in the C terminus of OprF had 76% sequence similarity to the C terminus of OmpA (22). However, the N-terminal β-barrel regions of these two proteins did not share sequence homology. At the time, OmpA was a controversial OM porin, as researchers debated whether this protein was a porin at all due to its slow solute diffusion rate, which is 2 orders of magnitude slower than that of other E. coli porins, such as OmpF (23, 24). In 1989, Woodruff and Hancock confirmed that substantial similarities exist between OprF and OmpA. OprF cross-reacted with polyclonal antibodies raised against OmpA, and cellular growth defects exhibited by an E. coli OmpA transposon mutant were restored by complementation with OprF (25). Similar to its P. aeruginosa counterpart, OmpA was eventually established as a major, nonspecific slow porin in E. coli (23). Additional OmpA homologs were also eventually found in other organisms. Today, the OmpA porin family is characterized by its N-terminal 8-stranded β-barrel OM domain and its globular, periplasmic C-terminal domain (26).

While OprF was originally characterized in 1979, it took another 20 years before the anomaly of its slow diffusion rate with high exclusion size was resolved. In 2000, the Hancock group established that the first 162 amino acids of the OprF N terminus formed a small, closed channel in lipid bilayers (27). This observation begged the question: how could a closed β-barrel be permeable to large solutes? Sugawara and Nikaido looked to the OprF homolog OmpA for answers. They had previously observed that OmpA had exhibited dual conformations with an “opened” and a “closed” form (23). The researchers examined OprF for similar folding patterns and found that during sedimentation of proteoliposomes, OprF, like OmpA, exists as both a “closed” conformer and an “open” conformer (Fig. 1), with 95% of the porin in the closed state (28).

FIG 1.

Speculative model of OprF functions in the biofilm. When cell associated, the majority of OprF (blue) is in the closed state (top, left), which is composed of an 8-stranded β-barrel OM domain (cylinder with thick wall) and a periplasmic peptidoglycan-interacting domain (bean shape). The closed conformer is predicted to be able to convert to an open state (top, right). The minor, open conformer is composed of a single β-barrel domain with 14 or more strands (cylinder with thin wall), lacks the peptidoglycan-interaction domain, and is predicted to oligomerize. We speculate that the two conformations and the interaction of OprF with LecB (tetramer of yellow ovals) may serve different roles in the P. aeruginosa (Pa) biofilm (green). The closed conformation may be involved in adhesion of cells to the host mucosa via a LecB bridge that interacts with both OprF and host surface glycoproteins (red hexagons) (A), cell-to-cell attachment in the biofilm via an OprF-LecB-OprF complex linking adjacent cells (B), and cell-to-matrix attachment via an OprF-LecB-Psl (orange hexagons) complex (C). (D) In comparison, the open conformation may be involved in the attachment of lipid vesicles (OMV) to the biofilm matrix via an OMV-OprF-LecB-Psl interaction complex.

The major conformation has two domains: an N-terminal 8-stranded β-barrel domain that spans the OM, which is linked to a C-terminal domain that associates with the peptidoglycan (Fig. 1, top left) (26, 29). This conformer represents the closed state of the porin. This closed conformation is generally considered necessary for cell envelope integrity and cell shape via its C-terminal peptidoglycan association (30). However, as noted by Rawling et al. (30), the exact role of the OprF closed conformer in OM stability needs to be further investigated. It was recently shown that in vivo, the C terminus of the closed OprF conformer interacts with OprI (a major OM lipoprotein) and OprL (a peptidoglycan-associated OM protein) in the periplasm via cross-linking studies (31). The association of these P. aeruginosa proteins is similar to that previously observed among their E. coli homologs OmpA, LPP, and Pal (32). Therefore, the interaction of multiple peptidoglycan-associated proteins is believed to contribute to the integrity of the OM, and there may be a cumulative and cooperative effect of multiple OM proteins in maintaining membrane stability, with OprF being only one component in OM stability.

In comparison to the closed major conformer, the minor conformation, which is the open state, accounts for 5% of the OprF found in the OM. It folds as a single β-barrel domain with 14 or more strands. This minor conformer can oligomerize into loose multiprotein complexes (Fig. 1, top right) (28). In this conformation, OprF does not associate with the peptidoglycan, and the high exclusion limit seen for OprF is attributed to the open oligomerized state, which has a 2-nm functional pore size (20). It has been proposed that in E. coli, the two-domain closed OmpA conformation represents a highly stable folding intermediate, with the open conformer representing the mature protein (33–35). Support for this intermediate-folding model has also been reported for OprF by Nestorovich et al. during their studies of channel conductivity (21). In a sample that was enriched for the open OprF conformation, these researchers reported variations in pore sizes, which they attributed to folding intermediates or “subconformations” (21).

While the majority of OmpA family members exist in the closed state in the OM under standard laboratory conditions, evidence suggests that regulation of the porin conformation is important for the physiology of the bacteria. In Salmonella enterica serovar Typhimurium, the OmpA homolog displays dynamic shifts in conformation in the presence of reactive oxygen species, such as hydrogen peroxide. When the bacteria are exposed to oxidative species, the S. enterica OmpA rapidly converts to the closed conformation, reducing the intake of oxidizing compounds and mitigating damage (36). Furthermore, in various Pseudomonas species, OprF displays conformational changes under temperature fluctuation. Cells at their optimal growth temperature had more OprF in the open conformation, while low temperatures result in more closed porins (37). It is unclear whether the benefit to the cell is due to the porin or the OM integrity function of OprF. Nonetheless, these results suggest that cells can respond to their environment by altering the conformation of their OmpA homologs. Further research is needed to characterize the folding patterns of OmpA family members under varied environmental conditions to determine the biologically relevant conformation of these porins.

COMMUNITY ASSOCIATION: OprF AND BIOFILMS

While understudied, reports describing the roles of OprF in biofilms are increasing (Table 1). Song et al. recently published that OprF is necessary for P. aeruginosa to sense surface stiffness during the attachment stage of biofilm formation (38). After attachment to a soft surface, P. aeruginosa PAO1 expresses significantly more of the secondary messenger c-di-GMP than cells on a stiff surface. High levels of c-di-GMP are associated with the upregulation of biofilm-related genes (39–41). An oprF transposon mutant exhibited no significant difference in c-di-GMP levels between the soft and stiff surfaces, indicating the oprF mutant lost mechanosensory capabilities. Interestingly, planktonic cells of the oprF mutant did produce significantly more c-di-GMP than wild-type cells, in agreement with an earlier study by Bouffartigues et al. (42). This study found that an oprF transposon mutant produces more c-di-GMP and biofilm biomass than wild-type P. aeruginosa (42). Wild-type biofilm formation is restored via the reduction in c-di-GMP levels by overproduction of a phosphodiesterase in the oprF mutant. The results of these studies, which were performed under aerobic conditions, conflict with the reduced biofilm phenotype observed by Yoon et al. from an oprF mutant under anaerobic growth conditions, indicating that oxygen availability may impact the effects of OprF on biofilm development (43). Other OmpA family members have also been reported to impact biofilm development. In addition to being abundant in E. coli biofilms (44), OmpA increases E. coli biofilm formation on hydrophobic surfaces by repressing cellulose production via induction of the Cpx envelope stress response pathway (45). In the nosocomial pathogen Acinetobacter baumannii, OmpA mutants exhibit deficiencies in biofilm production on abiotic surfaces and bacterial attachment to and aggregation on eukaryotic cells (46). In a recent review, Chevalier et al. suggest that these contrasting biofilm formation results may be due to the differences in growth conditions, such as oxygen availability and growth media (16) (Table 1). Altogether, these results show that OprF is involved in biofilm formation, and further study of cells under different growth conditions is needed to elucidate the temporal and condition-dependent effects of OprF on biofilm formation.

TABLE 1.

Comparison of biofilm phenotypes in OprF and OmpA mutants^a

Parameter	Condition(s) and result(s) for:
	OprF			Other OmpA homologs
	Song et al. (38)	Bouffartigues et al. (42)	Yoon et al. (43)	Ma and Wood (45)	Gaddy et al. (46)
Strain	P. aeruginosa PAO1	P. aeruginosa PAO1	P. aeruginosa PAO1	E. coli K-12	A. baumannii 19606
Mutation	oprF::Tn5	oprF::Ω	oprF mutation	ΔompA::Ω	ompA::Kmr
Static CV attachment assay	Aerobic; LB; PDMS surface; Increased attachment of mutant	NA	NA	NA	NA
Static CV biofilm assay	NA	Aerobic; LB; PS surface; increased biofilm in mutant	NA	Aerobic; LB; for PS, PP, and PVC surface, decreased biofilm in mutant; for glass surface, increased biofilm in mutant	Aerobic; LB; PS surface; decreased biofilm in mutant
Flow biofilm assay	NA	Aerobic; LB; glass surface; higher maximal thickness but weaker biovolume in mutant	Anaerobic; LBN; glass surface; decreased biofilm in mutant	NA	NA

^aCV, crystal violet; LB, lysogeny broth; LBN, lysogeny broth with 1% KNO₃; PDMS, polydimethylsiloxane; PS, polystyrene; PP, polypropylene; PVC, polyvinyl chloride; NA, assay was not performed in the study.

While the roles of the exopolysaccharides and eDNA in the biofilm are well described and lipid vesicle research is expanding, the roles of matrix proteins remain relatively understudied. To date, the functions of only a few proteins in the P. aeruginosa biofilm matrix have been characterized: CdrA in cell-to-cell adhesion and biofilm structure (47, 48), ecotin in defense against host neutrophil elastase (14), the Fap amyloid proteins in biofilm stiffness (49), and LecB in biofilm structure (50). However, a few proteomic analyses of the P. aeruginosa biofilm matrix have been published. While the reported putative matrix proteins vary, OprF has been identified in multiple of these biofilm matrix studies (12–14, 51). OprF has also been found to be in high abundance in OMVs (51, 52), which are a component of the biofilm matrix (53). While there is a relatively large breadth of data available on cell-associated OprF, OprF as a matrix protein has not been studied. The conformational state, the potential functions, and the potential interactions of OprF in the biofilm matrix are unknown.

As an OM component, OprF interacts with other molecules. Whether these interactions of OprF at the cell surface exist in the biofilm matrix as well is unknown. One example is the cell surface association of OprF with the lectin LecB (54), which binds fucose and mannose (55). LecB is a P. aeruginosa virulence factor and has been implicated in adhesion to host epithelial cells (56). In the absence of OprF, LecB dissociates from the OM and is found in the culture supernatant (54). We speculate that LecB can associate either directly or indirectly with an exposed extracellular loop of the closed OprF conformer, based on our back-of-the-envelope calculations from the work by Funken et al., who were able to visualize the amount of LecB bound to the surface of ∼2.4 × 10⁶ cells by immunoblotting (54). Since 95% of cell-associated OprF is in the closed form, there should be approximately 1.2 × 10⁶ molecules of LecB associated with each cell, assuming a one-to-one ratio of closed OprF with a LecB tetramer. There would, therefore, be ∼57 ng of LecB on the cells, which is detectable by immunoblotting, while 5% of that would not be detectable using the standard enhanced chemiluminescence reagent specified in the study. It is important to note, however, that this rationale does not preclude LecB binding with OprF in the open conformation.

Passos da Silva et al. recently reported that extracellular LecB bound to the P. aeruginosa matrix exopolysaccharide Psl stabilizes the biofilm (50). In light of the known association between OprF and LecB, an OprF-LecB-Psl tripartite interaction may exist in the biofilm matrix (Fig. 1C). Since LecB is a tetramer (57), it may interact with OprF via one subunit, while interacting with Psl via another, linking OprF to Psl. If this binding occurs when OprF is cell associated, we speculate that Psl can be tethered to bacterial cells via LecB, providing a stabilizing force to the matrix structure. Such a mechanism would also explain the role of LecB in retaining Psl in the biofilm, as seen by Passos da Silva et al. (50). Furthermore, since OprF is abundant in P. aeruginosa OMVs (51, 52), the potential for extracellular Psl-associated LecB to bind to OMV-associated OprF in the biofilm matrix exists. The potential for such a protein-mediated lipid-exopolysaccharide biofilm matrix complex, which has been previously proposed by Fong and Yildiz (58), may explain how OMVs are retained in the biofilm matrix (Fig. 1D). Further exploration of this macromatrix interaction and, in general, the function of matrix-associated OprF would greatly add to the current understanding of OprF roles in P. aeruginosa biofilms.

Extrapolating further, we reason that P. aeruginosa biofilms feature multiple permutations of this OprF-LecB-mediated binding (Fig. 1). LecB is localized in the cytoplasm in planktonic cells. It is only during the sessile lifestyle that LecB is translocated to the OM, suggesting that LecB is utilized during biofilm formation (56). Furthermore, lecB mutant strains produce deficient biofilms (50, 56). Due to the known interaction of cell-associated OprF with LecB, we speculate that this interaction contributes to P. aeruginosa cell aggregation in the biofilm (Fig. 1B). The cell-to-cell binding of two OprF proteins from neighboring cells to one LecB tetramer could promote aggregation and cell packing in the biofilm. In support of this aggregation hypothesis, OprF and LecB have been shown to be necessary for hemagglutination of red blood cells by P. aeruginosa via a putative bridging mechanism (54).

While OprF binding to LecB has been documented, a molecular understanding of this interaction remains unknown. Funken et al. hypothesized that LecB binds OprF via its sugar-binding site, since the interaction could be competed off via the addition of fucose (54). While this hypothesis assumes a competitive inhibition, it is possible that the disruption by fucose is noncompetitive. However, the structure of LecB is not greatly altered when fucose is bound (59), suggesting against a noncompetitive mechanism of inhibition by fucose on the OprF-LecB interaction. This elicits a question: why would the sugar-binding domain of LecB bind the protein OprF? The sugar-binding sites of LecB exhibit an unusually strong affinity for fucose, with a lessor affinity for mannose (55). Since the OprF-LecB interaction can be disrupted by the addition of fucose, not only does it point to LecB potentially interacting with OprF via its sugar-binding domain but it also suggests OprF may be fucosylated (54). While Funken et al. speculate that fucosylation of OprF may be responsible for LecB binding, they note that they were unable to obtain experimental evidence of OprF glycosylation. However, recent work by Khatua et al. identified OprF as a sialoglycoprotein via mass spectrometry when cells are grown on human serum (60), suggesting that OprF can be glycosylated. Further support for OprF glycosylation is the documented O-linked glycosylation of the OmpA-like Porphyromonas gingivalis protein PG0695 (61). Although these data suggest that OprF may also be glycosylated, it is worth noting that the P. gingivalis OmpA protein undergoes O-linked glycosylation of S271 (61). In an alignment of the two protein sequences, the P. gingivalis 15-amino-acid glycosylation region, which includes S271, is not well conserved in OprF. The OprF amino acid corresponding to this OmpA residue is an aspartic acid, D225. Since aspartic acid residues are not known to undergo O-linked glycosylation (62), it seems unlikely that P. aeruginosa OprF is glycosylated in the same fashion or location as the C-terminal glycosylation of P. gingivalis OmpA. However, since the P. gingivalis OmpA is glycosylated, it remains possible that OprF is as well, albeit at a different site.

THE TRAVELER: OprF AND LIPID VESICLES

OMVs are 50- to 300-nm spheres of OM that are commonly produced by Gram-negative bacteria and contain cargo designed for cell-to-cell communication, delivery of virulence factors, and evasion of host immune response (63). P. aeruginosa can produce OMVs from both planktonic and biofilm cells, but the vesicle quantity and protein content vary between lifestyles (53, 63). Abundant amounts of OprF have been seen in P. aeruginosa OMVs from both planktonic cell supernatants and biofilms (12, 51, 52).

While abundant in OMVs, OprF appears to inhibit OMV biogenesis, since an oprF transposon mutant produces 8-fold more OMVs than does the wild type (64). This phenotype is attributed to the role of OprF in regulating the production of the quorum-sensing molecule Pseudomonas quinolone signal (PQS), of which the mutant made 4-fold more and has been suggested to promote OMV biogenesis (64). How mutating OprF leads to this increase in PQS production is currently unclear. While studies have suggested that OprF plays a role in quorum-sensing regulation, the results are contradictory, with OprF mutants having various effects on PQS production under different conditions (64–66). However, the effect of OprF on OMV biogenesis is dependent on PQS production (64). Schertzer and Whiteley proposed that this is due to the bilayer-couple model for OMV biogenesis in P. aeruginosa, in which hydrophobic PQS inserts into the outer leaflet of the OM, asymmetrically expanding the membrane and leading to curvature. As the OM buds into a vesicle, periplasmic contents are pulled inside, and as PQS continues to accumulate in the outer leaflet, the OM pinches off to form an independent OMV (67). Recent work from the Schertzer group supports this model, showing that conditions promoting PQS production increase OMV production (68), that PQS can promote OMV formation in various species (69), and that PQS insertion induces membrane curvature in a computational model (70).

In contrast to the bilayer-couple model (67), a separate study by Turnbull et al. supports an alternative route to membrane vesicle (MV) biogenesis via prophage-mediated explosive cell lysis (71). Superresolution imaging captured lysed membrane fragments rapidly vesiculating. Unlike OMVs, which contain periplasmic content, MVs resulting from cell lysis encapsulate nucleic acids and cytoplasmic contents. A mutant lacking the prophage endolysin that is responsible for explosive cell lysis formed deficient biofilms (71). In biofilm matrices, MVs localize to the eDNA, where they reinforce the structural integrity of the biofilm (72). Although the OprF content of MVs is currently unknown, it is likely that the porin would be abundant in these vesicles formed from cell membrane shrapnel since it is abundant in the OM of intact cells. Furthermore, it is possible that both mechanisms of lipid vesicle formation occur in the biofilm, and the lipid vesicles in the biofilm encompass both OMVs with periplasmic content and MVs with cytoplasmic content.

It is tempting to speculate that there is an additional role for OprF in lipid vesicle formation. Since OprF in the closed state associates with the peptidoglycan but the open form does not, the conversion of OprF to the open state may promote OMV biogenesis by decreasing the envelope interactions. Supporting this speculation, in P. aeruginosa, the OM protein OprI inhibits OMV formation via its covalent bond with the peptidoglycan (64). Furthermore, in other organisms that do not produce PQS, it has been suggested that the increased production of OMVs by OmpA mutants is due to the decrease in cross-links between the OM and the peptidoglycan (63). These data lead us to believe that OMV-associated OprF exists in the open conformation, which lacks the peptidoglycan-interacting domain (Fig. 1D). With no peptidoglycan present in OMVs, we predict the C-terminal domain of OprF can fold into the membrane and oligomerize, creating the open, large pore. Furthermore, if the OMV-associated OprF open conformation can bind LecB, the previously proposed OMV-OprF-LecB-Psl matrix complex is possible and may be the mechanism by which OMVs are retained in the biofilm matrix (Fig. 1D).

INTRUDER ALERT: OprF IN VIRULENCE AND IMMUNE RESPONSE

OprF and other OmpA family porins are important for adhesion to host cells, which is a crucial step in bacterial pathogenesis (73). In P. aeruginosa, OprF is necessary for adhesion to human alveolar epithelial cells (65, 74). Attachment and virulence are significantly reduced during infection of human cells with oprF mutant strains (65). In addition, in E. coli-associated meningitis, OmpA is necessary for the bacterium to cross the blood-brain barrier and invade endothelial cells (75, 76). OmpA mutants exhibit a 25- to 50-fold decrease in brain endothelial cell invasion (75). Similarly, A. baumannii relies on OmpA for host epithelial cell invasion (77). In a murine lung model, A. baumannii OmpA mutants are attenuated for virulence relative to their highly virulent parental wild-type strains (77). In addition to its role in adhesion in A. baumannii, OmpA has been shown to play a role in cytotoxicity of this bacterium upon localization of the porin to the host cell nucleus and mitochondria (78). In addition to OprF, LecB has also been implicated in adhesion to host cells (54). We speculate that host cell adhesion may occur via a OprF-LecB complex, with LecB functioning as a sugar-binding bridge between cell-associated OprF and glycosylated human cells (Fig. 1A). Supporting this hypothesis, both OprF and LecB are necessary for aggregation of human erythrocytes by P. aeruginosa (54). If either protein is absent, the hemagglutination exhibited by wild-type P. aeruginosa is abrogated.

Given the role of OprF in cell attachment, it is unsurprising that the protein is highly immunogenic. During the immune response to P. aeruginosa infection, OprF is bound by C3b of the complement system, which tags the bacteria for phagocytosis by host macrophages and neutrophils (79). Furthermore, human interferon gamma (IFN-γ) also binds OprF (66, 80). P. aeruginosa does not, however, passively accept being targeted by the immune system. In response to the binding of IFN-γ to OprF, the bacterium upregulates a suite of virulence-associated genes (66). Fito-Boncompte et al. demonstrated a role for the porin in P. aeruginosa quorum sensing and virulence. An interruption mutant of OprF displayed disorganized quorum sensing, resulting in impaired biosynthesis of the extracellular virulence factors LasB, pyocyanin, LecA, and exotoxin A. From these results, these authors concluded that in wild-type P. aeruginosa, OprF may act as a sensor of the host immune system, modulating quorum-sensing signal production that then triggers a virulence response (65). Combined, these results suggest that OprF is necessary for P. aeruginosa to mount an effective virulence defense against the host immune response. It is worth noting that Fito-Boncompte et al. reported a delay, but not an increase, in PQS production in their oprF mutant. This finding contradicts the 4-fold PQS increase in the oprF mutant reported by Wessel et al., as discussed above (64). Wessel et al. attributed this contradiction to the differing PQS quantification methods used in the two studies (64). These authors contend that their approach of thin-layer chromatography provides more reliable PQS quantification than the liquid chromatography-mass spectrometry approach used by Fito-Boncompte et al. due to poor peak resolution of PQS during the liquid chromatography (64, 65). This contradiction highlights the need for further testing of OprF mutant strains and their virulence-related quorum-sensing response.

Neutrophil elastase is secreted by neutrophils at sites of P. aeruginosa infection (81). This serine protease is deployed by the innate immune system in order to degrade OprF, which results in the loss of bacterial OM integrity and P. aeruginosa cell death (82). The OprF homolog, OmpA, is also targeted by neutrophil elastase, resulting in the death of E. coli cells (83). In a fascinating potential defensive strategy, the presence of neutrophil elastase in the environment causes E. coli to secrete OmpA directly into the extracellular space and on the surface of OMVs (84). This secreted OmpA is believed to act as a decoy for neutrophil elastase, thereby reducing the interaction between the protease and its cell-associated target and protecting the cell from neutrophil elastase-mediated death. Whether this phenomenon also occurs in P. aeruginosa and in biofilms is unknown, but it is tempting to speculate that OMVs studded with OprF could sequester neutrophil elastase in the biofilm matrix, decreasing the abundance of the host protease and protecting the biofilm resident cells.

Due to the highly immunogenic nature of OprF, numerous studies have utilized the porin in pursuit of a P. aeruginosa vaccine (85, 86). Many recent publications involving the administration of chimeric OprF fusion proteins have demonstrated acquired immunity to P. aeruginosa can be achieved in mammals (87–91). Hassan et al. demonstrated that a recombinant trivalent OprF-OprI-flagellin vaccine significantly reduced mortality resulting from acute pneumonia in mice infected with mucoid and nonmucoid strains of P. aeruginosa (89). In addition, Gomi et al. showed that in a cystic fibrosis mouse model, mice with preexisting P. aeruginosa lung infections that were immunized with an OprF-adenoviral vector exhibited increased bacterial clearance relative to those that were not immunized (90). In contrast to these animal studies, Rello et al. demonstrated that while an OprF-OprI vaccine in human subjects undergoing ventilator therapy produced a significant immune response, there was no observed difference in P. aeruginosa infection rates between vaccinated and nonvaccinated groups (91). Nonetheless, OprF continues to show promise as a P. aeruginosa vaccine. Future phase II/III trials are needed to establish whether treatment with an OprF-based vaccine, and the subsequent immunogenic response, can protect humans against P. aeruginosa infections.

CONCLUDING REMARKS

While the cell-associated functions of OprF are well documented, additional research is necessary to elucidate the roles of OprF in the P. aeruginosa biofilm. Cell-associated OprF exists in two conformations: (i) a closed porin that provides structural support to the OM by anchoring it to the peptidoglycan, and (ii) an open β-barrel channel that is permeable to large, charged molecules (28). Given the multiple publications indicating that OprF plays a role in biofilm formation (42, 43) and is a matrix protein (12–14, 51), further research into its functions in the biofilm is needed.

We propose a conformation-based model for the functions of OprF in P. aeruginosa biofilms, OMVs, and virulence (Fig. 1). We hypothesize that OprF in the closed conformation, which is associated with the P. aeruginosa cell surface, interacts with the host epithelium (Fig. 1A), other P. aeruginosa cells (Fig. 1B), and the exopolysaccharide Psl (Fig. 1C) via a LecB bridge. Furthermore, we posit that in matrix-associated OMVs, OprF folds into the open conformation. If OprF in the open conformation also interacts with sugar-binding LecB, an OMV-OprF-LecB-Psl complex could form (Fig. 1D). It should be noted that while this model provides a framework for future OprF research, it is not necessarily exhaustive. In addition to the open and closed conformers, OprF may form oligomeric structures in the biofilm matrix, since purified OmpA has been reported to form dimers, trimers, and higher-order structures in vitro (92). Furthermore, folding intermediates between the defined open and closed conformations for OprF and OmpA have been documented (21), and these variations may have significant, unexpected effects on the OprF functions we have suggested in this model. Moreover, while we have focused on the potential roles of OprF and LecB interactions in the biofilm, OprF may be interacting with other matrix components that function outside the scope of our speculative model.

Ultimately, many questions remain unanswered. How do the environmental conditions influence OprF function in P. aeruginosa biofilm formation? Are the roles of open or closed OprF conformations in the biofilm matrix different than those of cell-associated OprF, and what is the influence of the OprF conformation on biofilm formation and virulence? What associations does OprF make with known matrix components? Probing of these questions will provide exciting new avenues of research and further insight into the physiology of P. aeruginosa biofilms.

Biographies

Erin K. Cassin is a graduate student in the laboratory of Boo Shan Tseng at the University of Nevada Las Vegas, Las Vegas, NV. Her thesis work focuses on the functions of matrix-associated proteins in P. aeruginosa biofilms.

Boo Shan Tseng is an assistant professor in the School of Life Sciences at the University of Nevada Las Vegas (UNLV), Las Vegas, NV. She received her Ph.D. with Hironori Funabiki at Rockefeller University in 2010 and was a postdoctoral fellow with Matthew R. Parsek at the University of Washington. She started her faculty position at UNLV in 2016, where she is continuing her work on the functions of the matrix in biofilm formation.