Role of Fraction 1 Antigen of Yersinia pestis in Inhibition of Phagocytosis

Abstract

Yersinia pestis, the causative agent of plague, expresses a capsule-like antigen, fraction 1 (F1), at 37°C. F1 is encoded by the caf1 gene located on the large 100-kb pFra plasmid, which is unique to Y. pestis. F1 is a surface polymer composed of a protein subunit, Caf1, with a molecular mass of 15.5 kDa. The secretion and assembly of F1 require the caf1M and caf1A genes, which are homologous to the chaperone and usher protein families required for biogenesis of pili. F1 has been implicated to be involved in the ability of Y. pestis to prevent uptake by macrophages. In this study we addressed the role of F1 antigen in inhibition of phagocytosis by the macrophage-like cell line J774. The Y. pestis strain EV76 was found to be highly resistant to uptake by J774 cells. An in-frame deletion of the caf1M gene of the Y. pestis strain EV76 was constructed and found to be unable to express F1 polymer on the bacterial surface. This strain had a somewhat lowered ability to prevent uptake by J774 cells. Strain EV76C, which is cured for the virulence plasmid common to the pathogenic Yersinia species, was, as expected, much reduced in its ability to resist uptake. A strain lacking both the virulence plasmid and caf1M was even further hampered in the ability to prevent uptake and, in this case, essentially all bacteria (95%) were phagocytosed. Thus, F1 and the virulence plasmid-encoded type III system act in concert to make Y. pestis highly resistant to uptake by phagocytes. In contrast to the type III effector proteins YopE and YopH, F1 did not have any influence on the general phagocytic ability of J774 cells. Expression of F1 also reduced the number of bacteria that interacted with the macrophages. This suggests that F1 prevents uptake by interfering at the level of receptor interaction in the phagocytosis process.

The genus Yersinia includes three pathogenic species: Yersinia pestis, Y. pseudotuberculosis, and Y. enterocolitica. Y. pestis infection causes a massive inflammatory response in affected lymph nodes, and the most common clinical scenario is referred to as bubonic plague, whereas Y. pseudotuberculosis and Y. enterocolitica cause self-limiting intestinal disease in humans (10). The pathogenic Yersinia species share a common virulence plasmid of ca. 70 kb in size that is essential for virulence (5, 20, 24, 25, 36, 37, 55). The virulence plasmids of Y. pestis and Y. pseudotuberculosis are very similar and functionally interchangeable (37, 38, 51). These virulence plasmids encode the type III secretion system, which serves to deliver Yop (Yersinia outer protein) virulence effector proteins into host cells. Two of these Yops, YopH and YopE, are particularly important for the ability of Yersinia to inhibit phagocytosis (39, 40). YopE has been demonstrated to function as a GTPase-activating protein to downregulate multiple Rho GTPases (6, 48), which leads to disruption of actin microfilaments in the target cell (40, 41). YopH is homologous to eukaryotic protein tyrosine phosphatases (PTPases) and is by far the most active of all known PTPases (28, 54). The presence of YopH is indispensable for the ability of the bacteria to block phagocytosis, as well as virulence, in a mouse infection model (19, 39). Early studies showed that YopH caused general dephosphorylation of the target cell phosphotyrosine proteins (8, 9, 27). In experiments with HeLa cells, YopH was found to interact with and dephosphorylate p130^Cas and focal adhesion kinase. Both of these proteins have been suggested to be specific substrates of YopH (7, 35). The YopH-dependent phagocytic inhibition involves blockage of a general phagocytic mechanism as phagocytes preexposed to YopH-expressing bacteria have a much-reduced ability to ingest other types of prey (19). In Y. pseudotuberculosis, YopH has also been shown to resist uptake via Fc receptors (immunoglobulin G [IgG] mediated). The Fc receptor-mediated phagocytosis is triggered by specific antibodies, which serve to link the foreign antigen to these receptors on the phagocyte (19). The function of YopE and YopH has mainly been studied in Y. pseudotuberculosis, but since the two species are closely related it is very likely that YopE and YopH have the same function in Y. pestis infections. Strains of Y. pestis not expressing YopE or YopH have also been found to be avirulent in a mouse infection model (47).

In addition to the virulence plasmid, Y. pestis has two additional plasmids, which are unique to Y. pestis (20). The smaller of these two plasmids, pPla, is ca. 9.5 kb in size and encodes the Pla protease. This protein exhibits coagulase activity at 30°C and can also activate plasminogen into plasmin at 37°C (4, 45). Pla has been suggested to be important for the ability of Y. pestis to disseminate from peripheral infection routes (subcutaneous or flea bite) and cause systemic infections (46). Recently, it was reported that Pla is important for the ability of Y. pestis to invade epithelial cells, such as HeLa cells (15). It is therefore possible that Pla can also serve as an adhesin or invasin for Y. pestis (15).

The large 100-kb plasmid, pFra encodes two potential virulence determinants that are unique to Y. pestis: murine toxin and the fraction 1 (F1) capsule-like antigen (13, 26). The murine toxin has been shown to have phospholipase D activity that relates to its toxicity to mice (30, 42). Similar to other capsules or capsule-like antigens, F1 has been suggested to be involved in the antiphagocytic activity reported for Y. pestis (12), but the contribution of F1 to this activity is not fully understood. The F1 antigen (15.5 kDa) forms a large gel-like capsule or envelope (3, 11, 21, 49). The capsule material is readily soluble and dissociates from the bacterium during in vitro cultivation. The structural genes for F1 (caf1) and the associated genes caf1M, calf1A, and caf1R have been cloned and sequenced. The structural gene for F1 has been shown to be homologous to interleukin-1β (IL-1β) and suggested to interact with IL-1 receptors (1). However, no data on the role of a potential F1-IL-1β interaction with Y. pestis interacting with host cells during infection have yet been obtained. The Caf1M protein shares homology with PapD, a chaperone protein required for assembly of pili in, for instance, Escherichia coli, and caf1M has been proposed to act as a chaperone for F1 with a role in posttranslational folding and secretion of F1. Molecular modelling of F1 and Caf1M predicts structures that are consistent with other chaperone systems (52, 53). Caf1A is an outer membrane protein with homology to PapC, which is involved in the assembly and anchoring of E. coli pilus structures (32). The 30-kDa Caf1R is a positive activator with homology to the AraC family of transcriptional activators (31). F1 antigen is an important protective antigen in Y. pestis. However, mutation of the F1 antigen gene does not significantly increase the 50% lethal dose in different animal models, but a somewhat prolonged survival of the infected animals has been reported (16, 17, 18, 50).

In this study, we wanted to investigate the role of F1 in blocking uptake by macrophages in relation to the established role of YopE and YopH in blocking signaling pathways essential for macrophage function. We show that mutants unable to express F1 on the surface are impaired in the ability to block uptake by macrophages, compared to an isogenic strain expressing F1. However, unlike the effect of YopE and YopH, F1 had no effect on the ability of the macrophage to phagocytose other prey, such as yeast particles. We also found that strains expressing F1 adhered less efficiently to macrophages. Based on our findings we suggest that F1 acts at the level of receptor interaction and somehow prevents binding between the pathogen and the phagocyte to occur in a manner that promotes phagocytosis.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains used in this study are listed in Table 1. The liquid growth medium for Y. pestis strains was brain heart infusion broth (Oxoid). The solid medium was blood agar base (Oxoid). E. coli strains were grown in Luria broth or on Luria agar. Carbenicillin and chloramphenicol were used at final concentrations of 100 and 25 μg/ml, respectively. For the induction of F1 antigen and Yop expression, overnight cultures grown at 23°C were diluted 20 times (to an optical density at 600 nm of ca. 0.05) in brain heart infusion broth with 2.5 mM CaCl₂, incubated an additional hour at 23°C, and then shifted to 37°C growth for an additional 5 h prior to infection of macrophages.

TABLE 1.

Y. pestis strains and plasmid used in this study

Strain	Genotype and relevant phenotype	Source or reference
EV76	Pgm-negative derivative of Y. pestis	R. Brubaker
EV76C	Virulence plasmid-cured derivative of EV76	R. Brubaker
EV76Δcaf1M	EV76 with in-frame deletion of caf1M	This study
EV76CΔcaf1M	EV76C with in-frame deletion of caf1M	This study
EV76Δcaf1M/pYD14	EV76Δcaf1M complemented with plasmid pYD14 expressing Caf1M	This study
EV76CΔcaf1M/pYD14	EV76CΔcaf1M complemented with plasmid pYD14 expressing Caf1M	This study
pYD14	9.3-kb EcoRI DNA fragment including f1 operon in pBR322 with same site	This study
pNQ705	Catr suicide vector	33

DNA methods and construction of mutants.

Preparations of plasmid DNA, restriction enzyme digestions, ligations, and transformations of E. coli were performed essentially as previously described by Sambrook et al. (43). DNA fragments were purified from an agarose gel by using Geneclean (Bio 101). Plasmids were introduced into Yersinia by electroporation as described by Conchas and Carniel (14). The operon encoding F1 was cloned from plasmid pFra as follows: purified pFra from EV76C was digested with EcoRI and ligated into similarly cleaved pBR322. The clones were screened by Western blotting with F1 antiserum. The resulting clone containing a 9.3-kb insert encoding the entire F1 operon, as well as genes encoding IS100 and T3-like ligase and an integrase-like gene, was denoted pYD14. The caf1M mutant strain containing an in-frame deletion of caf1M was constructed by an overlapping PCR method, followed by a double crossover mediated by a suicide plasmid pNQ705 (Table 1) as described by Galyov et al. (23). The primers used were CR2 (5′-GGTT GAAGCGACTCGAGATTCTCTTCTAT), CM3B (5′-TCTCGCGGGTTCGCAGCAAAACTAAGCATGCC), CM2 (5′-CTGCGAAC CCGCGAGATAAAGAGAGCCTAAAG), and CM1B (5′-GCTTGAGCTCTCATAAAGTCACATTTTTGGAATAC). The deletion spanned 264 bp of the caf1M gene from bp 65 to bp 329 in the coding sequence (whole gene, 774 bp) (22). The deletion in caf1M was verified by PCR and by Western blotting with antibodies raised against a glutathione S-transferase-Caf1M fusion protein (data not shown).

Phagocytosis assay.

The mouse macrophage-like cell line J774.1 (ATCC TIB 67) was seeded (2 × 10⁵ cells) on coverslips and grown to semiconfluence in Dulbecco minimal essential medium (DMEM) containing 10% heat-inactivated fetal calf serum. The cells were maintained at 37°C in a humidified atmosphere of 5% CO₂ in air. The macrophages were routinely grown in medium containing 5 μg of gentamicin/ml prior to infection with different Y. pestis strains.

Before the addition of bacteria, the cells grown on coverslips were washed three times with phosphate-buffered saline (PBS) plus 2.0 mM KCl (PBSA), and then DMEM containing 10% fetal calf serum without antibiotics was added. The macrophages were infected with bacteria at a ratio of ca. 20 CFU to each cell. Unless otherwise stated, contact between bacteria and macrophages was enhanced by low-speed centrifugation for 5 min at 400 × g. After 30 min of incubation at 37°C with 5% CO₂ in air, the coverslips were washed three times with PBSA and maintained at 4°C. Intra- and extracellularly located bacteria were distinguished by the double immunofluorescence method described previously (29, 39). Briefly, to stain extracellular bacteria, the coverslips with infected cells were washed and then incubated with rabbit anti-Yersinia antiserum raised against whole bacteria. Excess antiserum was removed by three washes in PBS. The coverslips were then fixed in ice-cold methanol for 90 s, dried, and subsequently incubated with TRITC (tetramethyl rhodamine isocyanate)-conjugated donkey anti-rabbit immunoglobulins (Jackson ImmunoResearch Laboratories, Inc.) for 30 min at 37°C. To stain all of the bacteria associated with the J774 cells, the coverslips were again incubated with anti-Yersinia serum for 1 h at 37°C, washed three times in PBS, and finally incubated with fluorescein isothiocyanate (FITC)-conjugated donkey anti-rabbit serum (Jackson ImmunoResearch Laboratories) for 1 h at 37°C. After three rinses three times in PBS the coverslips were mounted on a glass slide. The mounting medium consisted of 20% Airvol (Air Products and Chemicals, Utrecht, The Netherlands) and 4% Citifluor (Citifluor, Ltd., London, United Kingdom) in 20 mM Tris (pH 8.5). The specimens were examined in a fluorescence microscope (Axioscope; Carl Zeiss, Oberkochen, Germany) equipped with a Plan-apochromate 63×/1.40 oil immersion objective. The extracellular bacteria were examined by excitation at 530 to 585 nm, and the total cell-associated bacteria were detected at 450 to 490 nm. In each experiment, at least 100 randomly selected cells per coverslip were counted for extracellular and total bacteria.

Opsonization of bacteria with IgG.

The antiserum used in the opsonization experiments was raised in rabbits by immunization with whole bacteria of Y. pseudotuberculosis YPIII. The anti-Yersinia serum at a dilution of 1:250 was added to the bacterial culture during the last 30 min of the cultivation at 37°C prior to the infection of the macrophages.

Phagocytosis of yeast particles by J774.

The macrophages (J774.1) were first infected by Y. pestis pregrown at 37°C for 5 h at a ratio of 40:1 (bacterial CFU/cell) as described above. After 30 min of infection at 37°C, yeast particles were added at a ratio of 15:1 to J774 cells. The cells were allowed to phagocytose yeast particles for 30 min at 37°C, and then the coverslips were transferred to 4°C to interrupt phagocytosis. Extracellular yeast particles were stained by rabbit anti-Saccharomyces cerevisiae antiserum, followed by treatment with TRITC-conjugated donkey anti-rabbit immunoglobulins, and then fixed by the addition of cold methanol. Finally, the total yeast particles were stained by FITC-conjugated donkey anti-rabbit serum as described above for Y. pestis.

Assay of Y. pestis interaction with macrophages.

The macrophages were infected with different Y. pestis strains at 37°C both with or without the low-speed centrifugation step prior to infection. The infection ratio of bacteria to macrophages was 30 CFU to 1. Infection was stopped by fixing the samples in ice-cold methanol, and the number of bacteria associated with the macrophages was determined by immunofluorescence staining as described above. For each experiment, bacteria associated with at least 100 macrophages were counted in randomly selected fields.

Electron microscopy.

Bacteria were resuspended in 10 mM Tris-HCl buffer (pH 7.4) with 10 mM MgCl and allowed to adhere to Formvar-coated grids for 5 min at room temperature. The grids were incubated with monoclonal antibodies raised against purified F1 (kindly provided by Arthur Friedlander) for 10 min at room temperature. The grids were thoroughly rinsed with buffer and incubated with goat anti-mouse antibody conjugated with 10-nm gold particles (Biocell GAM10) diluted 10-fold. Finally, the grids were rinsed in distilled water and negatively stained with 1% sodium silicotungstate.

RESULTS

Caf1M is essential for surface expression of F1 antigen.

Since our aim was to study the role of F1 antigen in prevention of phagocytosis, we first decided to construct a mutant strain unable to express F1 antigen on the bacterial surface. Previous studies in which F1 had been expressed in E. coli had shown that Caf1M was essential for expression of F1 (22). We therefore constructed an isogenic mutant strain of Y. pestis from which the gene encoding the Caf1M chaperone was deleted. We first verified by Western blotting analysis that strain EV76Δcaf1M did not express the full-length Caf1M protein (data not shown). As expected, the caf1M mutant strain only expressed minute amounts of F1 protein when analyzed by immunoblotting of whole bacteria grown at 37°C. To verify that the chaperone mutant strain was unable to express any F1 antigen on the bacterial surface, we analyzed bacteria grown at 37°C with immunogold labeling techniques by using F1-specific antibodies. Strain EV76 was extensively labeled with gold particles (Fig. 1). The label extended out substantially from the bacterial surface, indicating that F1 is part of a polymeric structure. No staining was seen when strain EV76Δcaf1M was subjected to the same analysis (Fig. 1). Complementation of the chaperone mutant with plasmid pYD14 resulted in overexpression of F1 polymer, and in this case the staining extended even further out from the bacterium, forming a structure almost as wide as the bacterium. Thus, Caf1M is essential for secretion or surface exposure of F1, and a caf1M mutant can serve as a F1-negative strain in the studies of how F1 affects phagocytosis.

FIG. 1.

Immunoelectron microscopy of Y. pestis strains with F1 antibodies. The indicated strains of Y. pestis were grown at 37°C and were treated as described in Materials and Methods.

Effect of F1 on phagocytosis of Y. pestis in the absence of opsonizing antibodies.

To establish whether F1 had any influence on the phagocytosis of Y. pestis, different strains were analyzed in a phagocytosis assay by using the macrophage-like cell line J774.1. Strain EV76, which expresses both F1 and the plasmid-encoded type III secretion system, was highly resistant to uptake, and only ca. 5% of the bacteria were taken up (Fig. 2). However, this ability was strongly dependent on pregrowth of the bacteria at 37°C. Bacteria grown at room temperature were essentially unable to prevent uptake, and almost 100% were intracellular. When the isogenic caf1M mutant strain of EV76 was analyzed, a slightly increased uptake of bacteria was seen. In this case, ca. 30% of the bacteria were phagocytosed (Fig. 2). However, with the strain lacking the virulence plasmid-encoded type III secretion system, a more profound effect on phagocytosis was seen. In this case the number of internalized bacteria increased from ca. 5% for the wild-type strain EV76 to ca. 70% for the strain cured for the virulence plasmid (EV76C, Fig. 2). As expected, a strain lacking both F1 and a functional type III secretion system (EV76CΔcaf1M) was essentially unable to resist uptake, with only ca. 5% of the bacteria remaining extracellular. In control experiments, the caf1M mutant strains were complemented with plasmid pYD14 (EV76Δcaf1M/pYD14 and EV76CΔcaf1M/pYD14), and in these cases the F1-mediated blockage of phagocytosis was restored (Fig. 2).

FIG. 2.

Phagocytosis of different Y. pestis strains and Y. pseudotuberculosis by J774 cells. J774 cells were infected with different Y. pestis strains at a bacterium/cell ratio of 20:1 for 30 min at 37°C as described in Materials and Methods. All strains (except those labeled “23°C”) were grown 5 h at 37°C prior to the phagocytosis experiment. The diagram shows the average fraction (± the standard deviation from three experiments) of the bacteria that remained extracellular in the experiment.

Another important observation was that the effect of F1 on phagocytosis was seen only after 4 to 5 h of induction of F1 expression at 37°C prior to infection of the macrophages (data not shown). In contrast, previous work in Y. pseudotuberculosis has shown that only 30 min of induction at 37°C is required to see an effect of the type III system on uptake (39). This indicates that it is necessary for large amounts of F1 polymer to be present on the bacterial surface before an effect on phagocytosis can be seen. In summary, the relative contribution of F1 and the type III secretion system for strains pregrown 5 h at 37°C could be stated as follows: the loss of F1 resulted in a sixfold increase in uptake in a virulence plasmid-cured strain, and the loss of the virulence plasmid in an F1-negative strain resulted in a 14-fold increase in uptake; the loss of both F1 and the virulence plasmid resulted in an 18-fold increase in uptake. We conclude that F1 contributes to the ability of Y. pestis to block uptake by phagocytes and that F1 acts in concert with the type III secretion system, thereby making Y. pestis highly resistant to phagocytosis.

Role of F1 in phagocytosis in the presence of opsonizing antibodies.

YopH-mediated inhibition of phagocytosis by Y. pseudotuberculosis has been shown to occur also in the presence of IgG-opsonizing antibodies (Fc receptor mediated) (19). In order to elucidate whether F1 could also mediate resistance to Fc receptor-mediated phagocytosis, we adopted the methods used in the study by Fällman and coworkers and studied phagocytosis of Y. pestis in the presence of anti-Yersinia antibodies. The binding of anti-Yersinia antibodies (raised against Y. pseudotuberculosis, which does not express F1) to the Y. pestis strains was seen for all of the strains upon addition of FITC-conjugated anti-IgG (data not shown). For strain EV76 and the F1-negative strain (EV76Δcaf1M), no difference in uptake was seen in the presence of opsonizing antibodies (Fig. 3). For the strain cured for the virulence plasmid (EV76C), however, significantly higher numbers of bacteria were phagocytosed when the bacteria were opsonized than for nonopsonized bacteria. This strain expresses F1 but lacks the type III secretion system. This suggests that F1 cannot efficiently prevent Fc receptor-mediated phagocytosis in the absence of the type III secretion system. We conclude that, for Y. pestis, the type III secretion system also mediated translocation of YopE and YopH, making the bacterium resistant to Fc receptor-mediated phagocytosis and that F1 appears to contribute to block this type of uptake only in the presence of a functional type III secretion system.

FIG. 3.

Phagocytosis of IgG-opsonized Y. pestis strains by J774 cells. The experiment was performed as for Fig. 2. In this experiment, phagocytosis in the presence of opsonizing antibodies directed against whole bacteria (Y. pseudotuberculosis) was also included. Values for opsonized bacteria are indicated by black bars, and values for nonopsonized bacteria are indicated by shaded bars. The diagram shows the average fraction (± the standard deviation from three experiments) of the bacteria that remained extracellular in the experiment. For details, see Materials and Methods.

Effect of Y. pestis on phagocytosis of other prey by J774 cells.

To further study the nature of the phagocytic inhibition mediated by Y. pestis, the ability of the pathogen to inhibit uptake of an unrelated prey was investigated as described by Fällman et al. (19). J774 cells were preinfected with different strains of Y. pestis 30 min prior to exposure to IgG-opsonized yeast particles. The cells infected with strain EV76 showed a highly reduced capacity to ingest the yeast particles; only ca. 10% of all cell-associated yeast was intracellular compared to 60% for macrophages that were not preinfected with Y. pestis (Fig. 4). In contrast, macrophages infected with strains lacking the virulence plasmid were almost as efficient as uninfected macrophages at ingesting yeast particles. F1 antigen, however, had no significant effect on the ability of phagocytes to take up yeast particles (Fig. 4). These results show that, like that of Y. pseudotuberculosis, the type III secretion system of Y. pestis has a significant effect on the phagocytic ability of the J774 cells (19). The fact that expression of F1 antigen had no effect on macrophage function further suggests that the mechanism by which F1 prevents uptake is different from that of YopH.

FIG. 4.

Effect of Y. pestis on uptake of IgG-opsonized yeast particles by J774 cells. The macrophages were first preinfected with the indicated Y. pestis strains for 30 min at 37°C and then analyzed for the ability to ingest yeast particles added at a ratio of 15:1 (yeast particles/cell). The cells were allowed to phagocytose for 30 min, and the number of the particles ingested was then determined as described in Materials and Methods. The results shown represent the means ± the standard error of three separate experiments.

F1 expression reduces bacterial binding to J774 cells.

In order to study whether growth conditions and the expression of F1 antigen influenced the number of bacteria that interacted with the macrophages, it was necessary to omit the low-speed centrifugation step that promoted rapid contact between the bacteria and the J774 cells in the phagocytosis assay. In the standard phagocytosis assay the numbers of cell-associated bacteria remained similar for the different strains used (data not shown). However, when the macrophages were infected with different strains without forced contact by centrifugation, the numbers of cell-associated bacteria varied considerably for the different strains. For strain EV76 grown at room temperature, nearly 30% of the total number of bacteria added were associated with the macrophage (Fig. 5). If the same strain (EV76) was instead pregrown for 5 h at 37°C to induce high levels of F1 expression, only ca. 8% of the bacteria were associated with the macrophages. For all strains unable to express F1 on the surface grown at 37°C, the numbers of bacteria associated with the macrophage were significantly higher than for the corresponding wild-type strain (i.e., ca. 26% for EV76Δcaf1M and ca. 24% for EV76CΔcaf1M; Fig. 5). Complementation of the F1-negative strains with plasmid pYD14 resulted in near wild-type levels of macrophage-associated bacteria (Fig. 5). In contrast, there was no significant difference between the virulence plasmid-cured strain and the wild type, indicating that Yop expression did not affect binding to the macrophages. In conclusion, the expression of F1 prevented the interaction between Y. pestis and the macrophages.

FIG. 5.

Association of different Y. pestis with J774 cells. The cells were infected with bacteria (pregrown at 37°C for 5 h) without prior centrifugation to promote contact for 1 h at 37°C at a bacterium/cell ratio of 30:1. The number of bacteria associated with J774 cells was determined by fluorescent staining as described in Materials and Methods. The data given represent means ± the standard errors of the means of three separate experiments and are expressed as the percentage of total infected bacteria that were associated with J774 cells.

DISCUSSION

The classical work of Cavanaugh and Randall (12) showed that Y. pestis, once grown inside phagocytes, is able to resist uptake. The capsule-like F1 antigen was postulated to be involved in this ability. However, no studies of phagocytosis using isogenic F1-negative strains have been performed. We show here that F1 indeed contributes to the phagocytosis resistance of Y. pestis. In order to show a significant effect of F1 on phagocytosis, it was necessary to induce F1 expression for 4 to 5 h at 37°C prior to infecting the macrophages. It is possible that, during the intracellular passage of Y. pestis studied by Cavanaugh and Randall, F1 expression was induced since the infection was performed at 37°C. We have also noted that, if macrophages are infected with Y. pestis pregrown at room temperature, the bacteria are phagocytosed and F1 expression is induced by intracellular Y. pestis at 37°C (P. Cherepanov and Å. Forsberg, unpublished observations).

F1 is unique to Y. pestis and appears to give this pathogen a high capability to resist phagocytosis. In the assay system used here, <5% of strain EV76 was taken up, whereas for the strain unable to express F1 antigen ca. 30% was phagocytosed (Fig. 2). The importance of F1 antigen in Y. pestis infections is further supported by the fact that F1 is highly expressed during infection in animals, as well in humans, and can serve as a protective antigen (2, 34, 44). Therefore, it is somewhat surprising that mutants unable to express F1 are not highly attenuated in different animal infection models (16-18, 50). One major feature of the type III secretion system encoded by the virulence plasmid common to all pathogenic Yersinia species is the ability to block phagocytosis via the intracellular activity of the translocated effector proteins YopE and YopH (40). Mutations in either yopE or yopH render Yersinia spp., including Y. pestis, essentially avirulent in animal model infections (47). However, for this system, precultivation at 37°C to induce the system is also essential to promote phagocytosis resistance. Important in this case, an induction time of 30 min is sufficient. This implies that the type III secretion system is important during the early stages of infection, whereas F1 antigen once expressed may render Y. pestis even more able to resist uptake and to rapidly multiply extracellularly, leading to a lethal systemic infection. It is also conceivable that the type III secretion system of Y. pestis functions optimally only during early stages of infection since later on, when the surface of Y. pestis is covered with the F1 polymer, the contact-dependent delivery of Yop effectors may be less efficient.

The mechanism by which F1 blocks uptake is clearly different from that of YopE and YopH. YopH is an efficient PTPase that rapidly dephosphorylates key molecules in signaling pathways essential for phagocytosis (7, 35). This results in a general failure of macrophages to function in phagocytosis (19). F1, on the other hand, is not likely to be targeted into the host cell but appears to be a major surface polymer of Y. pestis (Fig. 1). Even though F1 contributes to phagocytosis resistance, our study shows that this occurs without any major impact on macrophage function, since the expression of F1 did not impair the ability of J774 cells to ingest yeast particles (Fig. 4). F1 antigen has been found to be homologous to IL-1β, and this has led to suggestions that F1 might serve as an adhesin for Y. pestis (1) and possibly also as the contact-dependent adhesin for the type III secretion system. However, we found in our macrophage infection model that the expression of F1 instead prevented the interaction between Y. pestis and the J774 cell and that the type III secretion system was clearly also functional in the absence of F1. A similar observation was recently made by Cowan and coworkers (15), who noted that there was a temperature-induced property of Y. pestis that inhibited invasion of epithelial cells. These investigators suggested that F1 could be responsible for this effect. Based on this finding, it is possible that the antiadhesive effect of F1 is more general and not only active against phagocytic cells. These observations lend support to a mechanism in which F1 serves to prevent Y. pestis from interacting at least with phagocytic cells by masking adhesin-receptor interactions that potentially could result in the uptake of the pathogen. This could also explain why F1 could partly block Fc receptor phagocytosis, since the F1 polymer could also partly prevent opsonizing antibodies from binding to the bacterial surface. However, F1 most likely did not completely prevent opsonizing antibodies from interacting with the bacterial surface, since the antibodies to Y. pseudotuberculosis still bound to F1-expressing bacteria in the immunofluorescence assay that were used in the phagocytosis assay (see Materials and Methods).

Another interesting question is the mechanism by which F1 can serve as a protective antigen. It is tempting to speculate that F1 antibodies can serve as opsonins and promote the phagocytosis of Y. pestis. It is possible that, if expression of F1 during infection is high, then the type III secretion system could fail to prevent the blockage of uptake via Fc receptors. This could be due to the steric hindrance imposed by the F1 polymer, which in turn could prevent the type III system from interacting with the phagocytes in a manner that allows translocation.

In summary, we have shown that the F1 antigen provides Y. pestis with an additional mechanism for blocking phagocytosis that works by a mechanism different from that of the type III secretion system. F1 also prevents the association with phagocytes, presumably by preventing adhesin-receptor interactions. Future studies should clarify whether high levels of F1 expression also interfere with the cell contact-dependent type III secretion system of Y. pestis.