Previous Article | Next Article ![]()
Journal of Bacteriology, May 2008, p. 3381-3385, Vol. 190, No. 9
0021-9193/08/$08.00+0 doi:10.1128/JB.01840-07
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

James B. Bliska, and
David G. Thanassi*
Center for Infectious Diseases, Department of Molecular Genetics and Microbiology, Stony Brook University, Stony Brook, New York 11794-5120
Received 21 November 2007/ Accepted 19 February 2008
|
|
|---|
caf1A mutant demonstrated that the Caf1A usher is required for the assembly and secretion of the fraction 1 capsule. The capsule assembled into thin fibrils and denser aggregates on the bacterial surface. Pilus-like fibers were also detected on the surface of Y. pestis. The capsule occasionally coated these fibers, suggesting how the capsule may cloak surface features to prevent host recognition. |
|
|---|
![]() View larger version (40K): [in a new window] |
FIG. 1. CU pathways of Y. pestis. (A) The CU gene clusters are depicted, with the caf (F1) and psa (pH 6) systems at the top and the novel CU pathways below them. Usher gene y1543 is disrupted by an insertion sequence. The usher for the y4060-y4063 pathway is disrupted by a frameshift mutation into two open reading frames (y4061 and y4062). (B) Alignment of the F1-G1 loop region of the Y. pestis chaperones together with the prototypical FGS chaperones FimC and PapD. Each of the novel Y. pestis chaperones has a short F1-G1 loop region, indicating that they belong to the FGS subfamily. In contrast, Caf1M and PsaB have a longer F1-G1 loop region.
|
caf1A deletion mutant of strain KIM6+ (29) using the lambda red recombination method (4, 5) and primers ycaf1A_kanF (TCTGGAAATATCGACTTCCGTCTAGAAAAACATAATGGAAAAGAACTTCTTTTGTAGGCTGGAGCTGCTTCG) and ycaf1A_kanR (CCTGGTACCGATTAAGGGTATTTTGCGAGACTGTTATTTGGACAAGGTAAACCAAGATATCAATGGTAA). Proper construction of the
caf1A mutant was confirmed by PCR (data not shown). We first examined KIM6+ and KIM6+
caf1A for F1 capsule assembly on the cell surface by immunofluorescence microscopy. Bacteria were grown in heart infusion (HI) broth at 37°C to logarithmic phase and adsorbed to poly-L-lysine-coated glass coverslips. The capsule was visualized by incubating the bacteria with a monoclonal anti-F1 antibody (RDI) followed by a tetramethyl rhodamine isothiocyanate-conjugated secondary antibody (Sigma). Whereas Caf1 was detected on the surface of KIM6+ bacteria, no capsule was detected on KIM6+
caf1A bacteria (Fig. 2A). To complement KIM6+
caf1A, we amplified caf1A from Y. pestis by PCR using primers CAFF3226 (TGATGAATTCAAAGGACTAGCGGGAGCACG) and CAFR547 (CGAACTGCAGCGTAGAGAGGGCTTGTGTCC). The amplicon was cloned into the pGEM-T Easy vector (Promega) and then subcloned using EcoRI into the expression vector pMMB66 (19), creating plasmid pCaf1A. The addition of pCaf1A to KIM6+
caf1A restored the expression of the capsule (Fig. 2A), confirming that the usher is required for the assembly of F1 on the bacterial surface.
![]() View larger version (32K): [in a new window] |
FIG. 2. The Caf1A usher is required for the assembly and secretion of the F1 capsule. (A) Phase-contrast and corresponding epifluorescence images of Y. pestis KIM6+, the KIM6+ caf1A mutant, and the mutant complemented with pCaf1A. The capsule was detected on intact bacteria using a monoclonal anti-F1 antibody followed by a tetramethyl rhodamine isocyanate-conjugated secondary antibody. Images were taken with a 40x objective. (B) Secretion of Caf1 into the culture medium by KIM6+ and KIM6+ caf1A. Supernatant fractions were precipitated with trichloroacetic acid, resuspended in sodium dodecyl sulfate sample buffer, and incubated at 25 or 95°C as indicated. Samples were separated by gel electrophoresis and immunoblotted with monoclonal (25°C) or polyclonal (95°C) F1 antibodies.
|
caf1A grown to logarithmic phase in TMH (pH 7.4) defined medium (26) at 37°C. The supernatant fractions were centrifuged to remove bacteria, passed through a 0.22-µm filter (Millipore), precipitated with 9% trichloroacetic acid, subjected to gel electrophoresis, and immunoblotted with anti-Caf1 antibodies. Caf1 subunits undergo spontaneous polymerization in the periplasm, particularly in the absence of the Caf1A usher (31, 32). Polymerized Caf1 is resistant to dissociation and can be detected by gel electrophoresis if samples are incubated in sodium dodecyl sulfate sample buffer at 25°C (32). The anti-F1 monoclonal antibody reacted strongly with polymerized Caf1 but did not react well with preparations treated at 95°C to generate the denatured monomer (data not shown). In contrast, an anti-F1 rabbit polyclonal antibody (generated using Caf1 protein purified under denaturing conditions from a recombinant E. coli strain) reacted strongly with the denatured monomer. Therefore, to ensure sensitive detection of both forms of the Caf1 subunits, we used the F1 monoclonal antibody to probe samples treated at 25°C and the F1 polyclonal antibody to probe samples treated at 95°C. As shown in Fig. 2B, secreted Caf1 subunits were readily detected in the medium from KIM6+ cultures. However, no Caf1 was detected in the medium from KIM6+
caf1A cultures (Fig. 2B). Identical results were obtained by immunoprecipitation of Caf1 from culture supernatant fractions using the monoclonal anti-F1 antibody (data not shown). These findings demonstrate that the Caf1A usher is required for both the assembly and secretion of Caf1 in Y. pestis, in agreement with the model established for the biogenesis of pilus-type fibers by the CU pathway. Our results differ from those of a previous study by Karlyshev and coworkers of recombinant E. coli strains, which found that the Caf1A usher was not required for F1 secretion (13). In fact, we also observed that the usher was not required for Caf1 secretion in recombinant E. coli strains (L. M. Runco and D. G. Thanassi, unpublished data). In additional experiments, we noted that whereas the level of Caf1 in the periplasm of Y. pestis decreased in the absence of the usher, periplasmic Caf1 levels increased in E. coli (Runco and Thanassi, unpublished). This buildup of Caf1 subunits in E. coli in the absence of the usher likely results in nonspecific release of Caf1 into the culture medium. This also implies that Y. pestis possesses a mechanism to sense and respond to the inappropriate accumulation of Caf1 subunits in the periplasm.
We examined the ultrastructure of the Y. pestis capsule by whole-bacterium, negative-stain transmission electron microscopy (transmission EM). Bacteria were grown in HI broth or TMH (pH 7.4) to logarithmic phase at 37°C, adsorbed to polyvinyl formal carbon-coated grids (Ernest F. Fullam, Inc.), fixed with 1% glutaraldehyde, washed successively with phosphate-buffered saline (PBS) and water, and then stained with 0.5% phosphotungstic acid (Ted Pella, Inc.). The appearance of the capsule was more distinct than reported in previous studies, in which the capsule generally appeared as an amorphous haze or dense mass surrounding the bacteria (3, 7, 17). KIM6+ grown in TMH consistently produced an extended halo composed of thin fibrils and interspersed denser aggregates (Fig. 3A). The
caf1A deletion mutant never expressed similar structures (Fig. 3E and G). Occasionally with Y. pestis grown in TMH, but more consistently with bacteria grown in HI broth, the capsule appeared as a dense structure that more closely resembled previously reported images (Fig. 3D) (3, 7, 17). This dense capsular material, likely composed of aggregates of the thin fibrils, sometimes extended out from the bacterial surface in long strands (Fig. 3D, H, and I). The thin, fibrillar appearance of the F1 capsule resembles structures previously reported for other members of the FGL subfamily of CU pathways, including the pH 6 antigen of Yersinia (11, 16) and the CS3 and CS6 pili of enterotoxigenic E. coli (14, 15). This suggests a common structure and assembly mechanism for members of the FGL subfamily.
![]() View larger version (138K): [in a new window] |
FIG. 3. Whole-bacterium, negative-stain transmission EM of Y. pestis. (A) KIM6+ grown in TMH, showing the F1 capsule as a fine, fibrillar network. (B) KIM6+ grown in TMH. The bacterium on the left is expressing multiple pilus-like fibers (black arrows), while the bacterium on the right is expressing capsular material. (C) KIM6+ grown in TMH, showing bacteria lacking the F1 capsule but expressing multiple pilus-like fibers. (D) KIM6+ grown in HI broth, showing the F1 capsule as a denser structure with extended strands. (E) KIM6+ caf1A grown in TMH. No surface structures are present. (F) Immunogold EM of KIM6+ grown in TMH and incubated with monoclonal F1 antibody. The capsular material is extensively labeled with gold particles. (G) Immunogold EM KIM6+ caf1A grown in TMH and incubated with the monoclonal F1 antibody. No gold labeling was detected despite the presence of pilus-like fibers. (H and I) Immunogold EM of KIM6+ grown in HI broth and incubated with monoclonal F1 antibody. Labeled capsular material (white arrows) is present along with unlabeled pilus-like fibers (black arrows). The capsule appears to coat some pilus fibers. Bars = 500 nm.
|
caf1A (Fig. 3G), and no gold labeling was detected when an isotype-matched control monoclonal antibody or the preimmune rabbit serum was used in place of the anti-F1 antibodies (data not shown). The pH 6 antigen (1), encoded by the psa gene cluster (Fig. 1A), should be repressed under our culture conditions, and examination of the Y. pestis
psa strain DSY13 (17) by EM revealed the same F1 structures as detected on KIM6+ (data not shown), confirming that the pH 6 antigen does not contribute to these structures.
In addition to bacteria being completely surrounded by the F1 capsule,
10 to 15% of the KIM6+ bacteria were only partly coated with capsular material or lacked the F1 capsule altogether (Fig. 3A to C). Notably, long, thin, pilus-like fibers were apparent and were most often observed on bacteria not expressing the F1 capsule (Fig. 3B and C). KIM6+
caf1A produced the same pilus-like fibers, which were not labeled by the anti-F1 antibody (Fig. 3G). The
psa strain DSY13 also produced similar fibers (data not shown), demonstrating that these fibers are not related to the capsule or the pH 6 antigen. In agreement with this, a recent study noted the presence of isolated, pilus-like fibers on the surface of a Y. pestis
caf
psa double mutant (17). Y. pestis contains eight putative CU gene clusters in addition to the caf and psa gene clusters (Fig. 1A) (21). Alignment of the chaperones from these novel CU gene clusters shows that they group with the prototypical FSG chaperones PapD and FimC in the F1-G1 strand region (Fig. 1B). Therefore, the novel CU gene clusters belong to the FGS subfamily that assembles pilus-type structures, suggesting that one or more of these pathways could encode the fibers observed by EM. Indeed, analysis of the novel Y. pestis CU gene clusters expressed in E. coli revealed that they are capable of assembling pili (8). The pilus fibers could serve as adhesins or invasins of Y. pestis, which lacks the proteins known to provide these functions in Yersinia pseudotuberculosis and Yersinia enterocolitica (18, 22, 25).
Interestingly, in KIM6+ strains expressing both the F1 capsule and the pilus-like fibers, the capsule sometimes appeared to coat the pilus-like fibers. This was particularly so for strains grown in HI broth, where the capsule adopted the more-condensed form, resulting in the capsule extending away from the bacterial surface in long strands (Fig. 3H and I). These extensions were clearly labeled by anti-F1-directed gold particles, while the pilus-type fibers were unlabeled or labeled by gold particles only where capsular material was present (Fig. 3H and I). Expression of the F1 capsule is known to inhibit the adhesion of Y. pestis to epithelial cells and to decrease uptake by macrophages (7, 17). The ability of the F1 capsule to cover even extended surface structures such as the pilus fibers suggests a mechanism whereby the capsule physically cloaks surface features of Y. pestis to prevent its recognition or uptake by host cells.
The Expression Core is supported by Northeast Biodefense Center grant U54-AI057158-Lipkin. This work was supported by National Institutes of Health grants AI055621 (to D.G.T. and J.B.B.) and U54-AI057158-Lipkin (to J.B.B.).
Published ahead of print on 29 February 2008. ![]()
Present address: Institute of Clinical Microbiology, Immunology and Hygiene, University Hospital Erlangen, Wasserturmstr. 3-5, D-91054 Erlangen, Germany. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»