The Salmonella enterica Serovar Typhi Type IVB Self-Association Pili A
http://www.100md.com
《感染与免疫杂志》
Department of Biochemistry, Hong Kong University of Science and Technology, Clear Water Bay, Hong Kong, China
ABSTRACT
Salmonella enterica serovar Typhi and some strains (Vi+) of serovar Dublin use type IVB pili to facilitate bacterial self-association, but only when the PilV proteins (potential minor pilus proteins) are not synthesized. Pilus-mediated self-association may be important in the pathogenesis of enteric fever. We have shown previously that the extent of DNA supercoiling controls the rate of Rci-catalyzed inversion of a DNA fragment which includes the C-terminal portions of the PilV proteins. This inversion therefore controls PilV synthesis as a high inversion rate prohibits transcription of pilV-encoding DNA. Here, we describe the manner in which PilV protein expression inhibits bacterial self-association and present data which suggest that incorporation of one or a few PilV protein molecules into a growing pilus, comprised of PilS subunits, causes the pilus to detach at the bacterial membrane. The bacteria are then unable to self-associate. We suggest that this phenomenon may be relevant to the pathogenesis of typhoid fever.
TEXT
Annually, 16 million cases of typhoid fever occur, and these cases are associated with 600,000 deaths (4, 12). While Salmonella enterica serovar Paratyphi strains account for some morbidity, S. enterica serovar Typhi causes much higher mortality than serovar Paratyphi does. We have suggested previously that pilus-mediated bacterial self-association and PilS-mediated bacterial attachment to the cystic fibrosis transmembrane conductance regulator (CFTR) may contribute to bacterial virulence (13, 17). It is possible that PilS acts as the protease-sensitive, heat-sensitive factor known to be required for CFTR recruitment (8). It is likely that an early PilS-CFTR interaction is replaced by a more intimate association between bacterial lipopolysaccharide and CFTR as the bacterium comes into closer contact with the mucosal cell wall (8). The association of PilS with CFTR may be important in facilitating the entry of serovar Typhi not only into intestinal epithelial cells (17) but also into human monocytes after bacterial penetration of the gut epithelium. Piliated serovar Typhi stimulated protein kinase C activity, interleukin 6 production, and nuclear transcription factor NF-B synthesis in human monocytic THP-1 cells to extents greater than those seen with the control pil insertion mutant (18). A 12-residue peptide that was selected, using a ribosome display system, to bind to PilS resulted in 75% inhibition of piliated serovar Typhi entry into human monocytic THP-1 cells when it was used at a concentration of 6 μM, while a control peptide had no effect; neither peptide inhibited THP-1 cell invasion by a nonpiliated strain of serovar Typhi (19). Also, a 71-nucleotide RNA aptamer, obtained by systematic evolution using exponential enrichment, inhibited (71 to 74%) serovar Typhi entry into THP-1 cells, although the same aptamer also nonspecifically inhibited (19 to 33%) of the THP-1 cell entry by a pilS mutant (11). The entry of piliated serovar Typhi into human monocytic THP-1 cells was reduced by only 50% when the piliated and nonpiliated strains were used in mixed infections but was reduced by 93 to 94% when the strains were examined singly (11). As piliated serovar Typhi may enmesh nonpiliated bacteria (9), it is likely that human cell entry tests using bacterial mixtures which include a positive control strain overestimate the invasive capacity of an admixed nonpiliated mutant.
Using a model in which self-association of serovar Typhi strains results in greater transfer to such strains of a transmissible plasmid from an enmeshed donor strain, we showed that bacterial self-association occurs only when the presumptive minor pilus proteins PilV1 and PilV2 are not expressed (9). DNA inversion activity of a shufflon controls the synthesis of the PilV proteins (6, 7, 9, 16). Slow shufflon inversion permits PilV synthesis. Inactivation of the shufflon by mutation renders PilV synthesis constitutive. Of particular interest is why constitutive expression of PilV proteins results in elimination of the ability of serovar Typhi cells to self-associate (9). Such cells seem to be competent for pilus production, as wild-type levels of PilS protein may be found in immunoblots of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) gels of material from the supernatants of cultures grown under shaking conditions (9).
Previously, we tried to increase pilus expression by laboratory-grown serovar Typhi. Insertion of a tac promoter upstream of pil did not enhance pilus production, possibly because overproduction of PilL (an outer membrane protein [14]) and/or PilM (an inner membrane protein [14]) destabilized bacterial membrane function. We then placed the pilN-pilV genes of serovar Typhi J341 under tac promoter control (21) while we allowed the proximal pil genes pilL and pilM to remain under pil promoter control. We constructed mutants with alterations in the pilS, pilV, or rci genes (9). In the wild-type strain an active Rci gene product may invert shufflon DNA to place either of two distinct C termini on a constant N-terminal pilV region, thus creating two distinct pilV genes designated pilV1 and pilV2. The pilV mutant does not express any PilV protein. Mutants in which rci was inactivated expressed either the pilV1 or pilV2 gene constitutively. A pilS mutation has been described previously (21). Notably, all strains except the pilS control strain expressed PilS protein in the supernatants of cultures grown (with shaking) in a manner that permitted shearing of pili (9). While a pilV mutant and a wild-type strain (both lacking the tac promoter) self-associated and a pilS mutant did not self-associate, it was noted that pilV1(Con) and pilV2(Con) mutants (lacking the tac promoter; formerly published as pilV1C and pilV2C) were also incapable of self-association, even though such strains seemed perfectly able to synthesize pili. To explore this apparent anomaly, the serovar Typhi strains with the pilN-pilV genes under tac promoter control were grown in static cultures for 48 h and examined by electron microscopy (Fig. 1). A strain with a wild-type shufflon region was heavily piliated (Fig. 1A), as was the pilV mutant (Fig. 1B). We prepared grids from each of five different cultures of both of the strains with the wild-type shufflon and the strain with the pilV mutation and examined 100 bacteria on each grid. We found that 53% of the wild-type bacteria were piliated, as were 90% of the organisms harboring the pilV mutation. This suggested that PilV expression inhibited pilus production. We measured the lengths of the pili on each of 100 bacteria on each grid and calculated that 14% of wild-type pili were >300 nm long, while the corresponding value for pili synthesized by the pilV mutant was 33%. The wild-type pili were therefore shorter on average than the pili synthesized by the pilV mutant. As expected, a pilS mutant lacked pili (Fig. 1C). Notably, the strains bearing either the pilV1(Con) or pilV2(Con) mutation were also nonpiliated (Fig. 1D and E). We used immunogold labeling to show that the pili of the strains with either the wild-type shufflon or the pilV gene were assembled from PilS subunits (Fig. 2A and B). We detected pilus tip PilV protein by immunogold labeling (Fig. 2C) on only a minority (<1%) of wild-type pili. We discuss this finding below. Exposure of grids to preimmune sera did not permit subsequent gold labeling of pilus components by the secondary antibody.
It is important to ask if strains bearing the tac promoter before pilN produce normally assembled pili. The micrographs in Fig. 1A and 1B show the production of recognizable pili under such conditions. The tac insertion should result in strong transcription enhancement of (among other pil genes) both pilS and pilV. There is no reason to suggest that the ratio of pilS transcription to pilV transcription in the strains used here is different from the ratio in a wild-type strain. When strains with normal pil operons were used in self-association tests (9), it was clear that PilV expression eliminated self-association under conditions under which self-association was measurable with wild-type bacteria or a pilV mutant. This may be explained if PilV production eliminates pilus assembly in wild-type serovar Typhi, as was found in the experiments using tac-driven pilN-pilV described here.
As the strains bearing either the pilV1(Con) or pilV2(Con) mutation were nonpiliated but expressed apparently normal (wild-type) amounts of PilS protein in the supernatants of cells grown in shaking cultures (9), it seemed possible that constitutive expression of a PilV protein resulted in pilus detachment even when cells were grown in the absence of shearing forces. We wished to minimize growth conditions that might shear pili, but growth in static cultures was very slow; the stationary phase was entered by 1-liter cultures in 2.8-liter flasks only after ca. 72 h of incubation at 37°C. We found that shaking growth conditions (i) permitted the strains to reach the stationary phase after ca. 20 h of growth but (ii) did not seem to cause pilus shearing provided that the depth of the culture medium was such that simple swirling without obvious agitation of the culture was observed. We therefore prepared 1-liter cultures in 2-liter flasks. The medium depth was 6.2 cm in these flasks. In our previous work (9) 250-ml cultures were grown in 1-liter flasks with a culture depth of 2.2 cm and associated agitation during shaking growth. Small volumes (20 ml) of stationary-phase cultures were inoculated into 1 liter of fresh LB medium and shaken (220 rpm) at 37°C for 20 h. The optical densities of the cultures were standardized, and the cultures were centrifuged at 5,000 x g for 20 min at 4°C. The supernatants were then centrifuged at 13,000 x g for 10 min at 4°C to remove all bacterial cells. The supernatants from this step were spun for 1 h in a Sorvall A-641 fixed-angle ultracentrifuge rotor operating at 186,453 x g and 4°C. The pelleted material from 1 liter of supernatant was resuspended in 600 μl of phosphate-buffered saline (pH 7.4), and this preparation was designated the crude pilus fraction.
To prepare a linear sucrose gradient, sucrose solutions containing 20, 30, 40, 50, and 60% (wt/wt) sucrose in phosphate-buffered saline (pH 7.4) were sequentially layered (2 ml each) to form a 10-ml step gradient, which was then allowed to diffuse for 14 h at 4°C. The crude pilus fraction (from ca. 1012 bacterial cells) was loaded and subjected to 45 min of centrifugation in a Sorvall TH-641 swinging-bucket rotor at 106,394 x g and 19°C. Fourteen fractions that were the same volume were collected from the top of the gradient. Aliquots (15 μl) of each fraction were analyzed by SDS-PAGE, followed by Western blotting using either mouse anti-prePilS (1:10,000) (11) or mouse anti-His6-prePilV1 (1:5,000) as the primary antiserum (Fig. 3). The supernatants of strains either with the wild-type shufflon (Fig. 3A) or with the pilV gene (Fig. 3B) contained very few pili. Traces of PilS protein were found in fractions 6 to 11 of the gradients, which presumably reflected minor shearing of full-length pili during the initial low-speed centrifugation step. A small amount of PilV protein was found in the culture supernatant of the strain with the wild-type shufflon (Fig. 3C). The PilV protein did not appear to be preferentially associated with the pili separated on the same gradient but rather remained in the top fraction, perhaps reflecting facile removal of PilV proteins from pilus tips, as discussed below. Some PilV protein was also found in the bottom gradient fractions. This may have represented PilV protein in bacterial membrane fragments in which the protein may have resided for a time prior to secretion. No PilV protein was found in culture supernatant from the strain carrying the pilV mutation (Fig. 3D). As expected, the pilS mutant strain expressed neither the PilS (Fig. 3E) nor PilV (Fig. 3F) protein. The pilS mutation, which resulted from insertion of a Kmr cassette into the pilS gene, should be polar on expression of the downstream pilV gene.
Notably, culture supernatants of strains carrying either the pilV1(Con) or pilV2(Con) mutation were rich in both the PilS (Fig. 3G and H) and PilV (Fig. 3J and K) proteins. In either case, the PilS proteins were distributed fairly evenly across gradient fractions, showing that pili of various sizes were first synthesized by cells of these strains and then detached, even when the cells were grown under conditions that minimized shearing of pili. As the PilV1 and PilV2 proteins are constitutively expressed in these strains, it was not surprising (given the secretion of PilV proteins by the strain with the wild-type shufflon, as mentioned above) to find that PilV proteins were prominent in culture supernatants. As previously noted in the case of PilV expression by the strain with the wild-type shufflon, the PilV1 and PilV2 proteins seemed to be enriched (i) in the top (less dense) gradient fractions and (ii) in the bottom (more dense) gradient fractions, where sizable membrane fragments containing PilV proteins which had not yet been exported would be expected. With the pilV1(Con) strain, both prePilV1 and PilV1 were found in the membrane-enriched (bottom) fractions, as double bands reactive with the anti-His6-prePilV antiserum appeared.
It is very likely that the PilV proteins are normally located at pilus tips because strains bearing various possible R64-encoded PilV proteins discriminate between bacterial recipients in liquid mating preparations (3, 20). The PilV proteins recognize determinants in the lipopolysaccharide molecules of the recipients, thus establishing connections with the donor (the R64-bearing strain) to facilitate transfer of the R64 plasmid. The PilV proteins do not seem to be stably pilus associated, however. With the strain expressing wild-type shufflon, however, the PilV proteins were not in the pilus-containing fractions. It is possible that the manipulations (high-speed centrifugation and pellet resuspension) required for preparation of material for sucrose gradient velocity centrifugation tend to result in detachment of PilV proteins from pilus tips. If this is true, then the favored localization of secreted PilV protein in the first gradient fraction [in both wild-type and pilV1(Con) and pilV2(Con) mutant strains] is explained. With the pilV1(Con) and pilV2(Con) mutant strains, smaller amounts of PilV proteins may be dispersed through the gradients. It seems likely that the high-frequency detachment of pili in strains constitutively expressing PilV proteins leads to high levels of PilV-tipped pili in the culture supernatants, and not all the PilV protein molecules are detached from the pili during manipulations prior to sucrose gradient velocity sedimentation. The suggested ease of detachment of PilV proteins from pilus tips may explain why it was difficult to find PilV-tipped pili by electron microscopy (Fig. 2C). While some long pili separated on the sucrose gradient may have retained pilus tip PilV proteins (Fig. 3J and K), immunogold microscopy of sucrose gradient fractions did not consistently reveal pilus tip PilV molecules, as manipulations prior to electron microscopy probably removed the already small amounts of pilus tip PilV proteins that remained pilus associated during sucrose gradient sedimentation. The data do not permit rigorous exclusion of the possibility that PilV secreted independently into culture medium acts externally to fracture pili containing exclusively PilS protein. If this occurred, however, then all fracture points would be at the outer membrane surface in pilV1(Con) and pilV2(Con) mutant strains (Fig. 1D and E) but remote from the membrane surface in the wild-type strain (Fig. 1A). We consider this unlikely.
The data described above allow us to propose that incorporation of a PilV protein into pilus growth at a site previously processing only PilS molecules causes detachment of the preexisting pilus. In serovar Typhi, pilus growth need not commence with a PilV protein, as the strain with the pilV gene is heavily piliated. In the strain with the wild-type shufflon a small amount of PilV protein appears in culture supernatants, compared to the levels seen in culture supernatants from strains constitutively expressing PilV1 and PilV2 proteins (Fig. 3, compare panel C with panels J and K). As we have previously shown, PilV protein expression is inhibited when shufflon inversion inhibits through-transcription of the pilV gene (16). The average pilus length of the strain with a wild-type shufflon was less than average pilus length of the pilV mutant (see above). This suggests that a small amount of PilV protein synthesis by the wild-type strain may indeed be involved in some pilus detachment. There were somewhat more pili in the culture supernatant of the strain with the wild-type shufflon compared with the very small amount of pili found in the culture supernatant of the strain carrying the pilV mutation (Fig. 3, compare panels A and B).
The serovar Typhi (and serovar Dublin and Paratyphi C [10, 15]) pil operons are clearly derived from a pil operon that is similar or identical to operons found on the enterobacterial plasmids R64 and R721. The shufflons of these plasmids are more complicated than that of serovar Typhi, and they have five or six 19-bp sequences, respectively, which serve as substrates for the rci-encoded recombinase (1, 2, 5). There is no evidence that serovar Typhi is a donor of DNA. Instead, we have suggested previously that the reduction in shufflon complexity in serovar Typhi to a simple two-state arrangement exploits the ability of PilV proteins to control pilus expression (9).
ACKNOWLEDGMENTS
This work was supported by the Hong Kong Government Research Grants Council.
We thank Tze Kin Cheung of the Materials Characterization and Preparation Facility of our university for his assistance with electron microscopy.
FOOTNOTES
Corresponding author. Present address: P.O. Box 20, Redlynch, Cairns, QLD 4870, Australia. Phone: (617) 4039-0939. E-mail: jhackett@cairns.net.au.
Present address: School of Optometry, University of California, Berkeley, CA 94720-2020.
Contact Christina Morris for strains or biological materials. E-mail: bctina@ust.hk.
REFERENCES
1. Gyohda, A., and T. Komano. 2000. Purification and characterization of the R64 shufflon-specific recombinase. J. Bacteriol. 182:2787-2792.
2. Gyohda, A., N. Furuya, N. Kogure, and T. Komano. 2002. Sequence-specific and non-specific binding of the Rci protein to the asymmetric recombination sites of the R64 shufflon. J. Mol. Biol. 318:975-983.
3. Ishiwa, A., and T. Komano. 2003. The pilus PilV adhesins of plasmid R64 recognize specific structures of the lipopolysaccharide molecules of recipient cells. J. Bacteriol. 185:5192-5199.
4. Ivanoff, B. 1995. Typhoid fever: global situation and WHO recommendations. Southeast Asian J. Trop. Med. Public Health 26:1-6.
5. Kim, S.-R., and T. Komano. 1992. Nucleotide sequence of the R721 shufflon. J. Bacteriol. 174:7053-7058.
6. Komano, T., S.-R. Kim, T. Yoshida, and T. Nisioka. 1994. DNA rearrangement of the shufflon determines recipient specificity in liquid mating of IncI1 plasmid R64. J. Mol. Biol. 243:6-9.
7. Kubo, A., A. Kusukawa, and T. Komano. 1988. Nucleotide sequence of the rci gene encoding shufflon-specific DNA recombinase in the IncI1 plasmid R64: homology to the site-specific recombinases of integrase family. Mol. Gen. Genet. 213:30-35.
8. Lyczak, J. B., and G. B. Pier. 2002. Salmonella enterica serovar Typhi modulates cell surface expression of its receptor, the cystic fibrosis transmembrane conductance regulator, on the intestinal epithelium. Infect. Immun. 70:6416-6423.
9. Morris, C., I. S. M. Tsui, C. M. C. Yip, D. K.-H. Wong, and J. Hackett. 2003. The shufflon of Salmonella enterica serovar Typhi regulates type IVB pilus-mediated bacterial self-association. Infect. Immun. 71:1141-1146.
10. Morris, C., C. K. P. Tam, T. S. P. Wallis, P. W. Jones, and J. Hackett. 2003. Salmonella enterica serovar Dublin strains which are Vi antigen-positive use type IVB pili for bacterial self-association and intestinal cell entry. Microb. Pathog. 35:279-284.
11. Pan, Q., X.-L. Zhang, H.-Y. Wu, P.-W. He, F. Wang, M.-S. Zhang, J.-M. Hu, B. Xia, and J. Wu. 2005. Aptamers that preferentially bind type IVB pili and inhibit human monocytic-cell invasion by Salmonella enterica serovar Typhi. Antimicrob. Agents Chemother. 49:4052-4060.
12. Parry, C. M., T. T. Hien, G. Dougan, N. J. White, and J. J. Farrar. 2002. Typhoid fever. N. Engl. J. Med. 347:1770-1782.
13. Pier, G. B., M. Grout, T. Zaidi, G. Meluleni, S. S. Mueschenborn, G. Banting, R. Ratcliff, M. J. Evans, and W. H. Colledge. 1998. Salmonella typhi uses CFTR to enter intestinal epithelial cells. Nature 393:79-82.
14. Sakai, D., and T. Komano. 2002. Genes required for plasmid R64 thin-pilus biogenesis: identification and localization of products of the pilK, pilM, pilP, pilR, and pilT genes. J. Bacteriol. 184:444-451.
15. Tam, C. K. P., J. Hackett, and C. Morris. 2004. Salmonella enterica serovar Paratyphi C carries an inactive shufflon. Infect. Immun. 72:22-28.
16. Tam, C. K. P., J. Hackett, and C. Morris. 2005. Rate of inversion of the Salmonella enterica shufflon regulates expression of invertible DNA. Infect. Immun. 73:5568-5577.
17. Tsui, I. S. M., C. M. C. Yip, J. Hackett, and C. Morris. 2003. The type IVB pili of Salmonella enterica serovar Typhi bind to the cystic fibrosis transmembrane conductance regulator. Infect. Immun. 71:6049-6050.
18. Wang, F., X.-L. Zhang, Y. Zhou, L. Ye, Z. Qi, and J. Wu. 2005. Type IVB piliated Salmonella typhi enhance IL-6 and NF-B production in human monocytic THP-1 cells through activation of protein kinase C. Immunobiology 210:283-293.
19. Wu, H.-Y., X.-L. Zhang, Q. Pan, and J. Wu. 2005. Functional selection of a type IV pilus-binding peptide that specifically inhibits Salmonella Typhi adhesion to/invasion of human monocytic cells. Peptides 26:2057-2063.
20. Yoshida, T., N. Furuya, M. Ishikura, T. Isobe, K. Haino-Fukushima, T. Ogawa, and T. Komano. 1998. Purification and characterization of thin pili of IncI1 plasmids ColIb-P9 and R64: formation of PilV-specific aggregates by type IV pili. J. Bacteriol. 180:2842-2848.
21. Zhang, X.-L., I. S. M. Tsui, C. M. C. Yip, A. W. Y. Fung, D. K.-H. Wong, X. Dai, Y. Yang, J. Hackett, and C. Morris. 2000. Salmonella enterica serovar Typhi uses type IVB pili to enter human intestinal epithelial cells. Infect. Immun. 68:3067-3073.(Connie K. P. Tam, Christina Morris, and )
ABSTRACT
Salmonella enterica serovar Typhi and some strains (Vi+) of serovar Dublin use type IVB pili to facilitate bacterial self-association, but only when the PilV proteins (potential minor pilus proteins) are not synthesized. Pilus-mediated self-association may be important in the pathogenesis of enteric fever. We have shown previously that the extent of DNA supercoiling controls the rate of Rci-catalyzed inversion of a DNA fragment which includes the C-terminal portions of the PilV proteins. This inversion therefore controls PilV synthesis as a high inversion rate prohibits transcription of pilV-encoding DNA. Here, we describe the manner in which PilV protein expression inhibits bacterial self-association and present data which suggest that incorporation of one or a few PilV protein molecules into a growing pilus, comprised of PilS subunits, causes the pilus to detach at the bacterial membrane. The bacteria are then unable to self-associate. We suggest that this phenomenon may be relevant to the pathogenesis of typhoid fever.
TEXT
Annually, 16 million cases of typhoid fever occur, and these cases are associated with 600,000 deaths (4, 12). While Salmonella enterica serovar Paratyphi strains account for some morbidity, S. enterica serovar Typhi causes much higher mortality than serovar Paratyphi does. We have suggested previously that pilus-mediated bacterial self-association and PilS-mediated bacterial attachment to the cystic fibrosis transmembrane conductance regulator (CFTR) may contribute to bacterial virulence (13, 17). It is possible that PilS acts as the protease-sensitive, heat-sensitive factor known to be required for CFTR recruitment (8). It is likely that an early PilS-CFTR interaction is replaced by a more intimate association between bacterial lipopolysaccharide and CFTR as the bacterium comes into closer contact with the mucosal cell wall (8). The association of PilS with CFTR may be important in facilitating the entry of serovar Typhi not only into intestinal epithelial cells (17) but also into human monocytes after bacterial penetration of the gut epithelium. Piliated serovar Typhi stimulated protein kinase C activity, interleukin 6 production, and nuclear transcription factor NF-B synthesis in human monocytic THP-1 cells to extents greater than those seen with the control pil insertion mutant (18). A 12-residue peptide that was selected, using a ribosome display system, to bind to PilS resulted in 75% inhibition of piliated serovar Typhi entry into human monocytic THP-1 cells when it was used at a concentration of 6 μM, while a control peptide had no effect; neither peptide inhibited THP-1 cell invasion by a nonpiliated strain of serovar Typhi (19). Also, a 71-nucleotide RNA aptamer, obtained by systematic evolution using exponential enrichment, inhibited (71 to 74%) serovar Typhi entry into THP-1 cells, although the same aptamer also nonspecifically inhibited (19 to 33%) of the THP-1 cell entry by a pilS mutant (11). The entry of piliated serovar Typhi into human monocytic THP-1 cells was reduced by only 50% when the piliated and nonpiliated strains were used in mixed infections but was reduced by 93 to 94% when the strains were examined singly (11). As piliated serovar Typhi may enmesh nonpiliated bacteria (9), it is likely that human cell entry tests using bacterial mixtures which include a positive control strain overestimate the invasive capacity of an admixed nonpiliated mutant.
Using a model in which self-association of serovar Typhi strains results in greater transfer to such strains of a transmissible plasmid from an enmeshed donor strain, we showed that bacterial self-association occurs only when the presumptive minor pilus proteins PilV1 and PilV2 are not expressed (9). DNA inversion activity of a shufflon controls the synthesis of the PilV proteins (6, 7, 9, 16). Slow shufflon inversion permits PilV synthesis. Inactivation of the shufflon by mutation renders PilV synthesis constitutive. Of particular interest is why constitutive expression of PilV proteins results in elimination of the ability of serovar Typhi cells to self-associate (9). Such cells seem to be competent for pilus production, as wild-type levels of PilS protein may be found in immunoblots of sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis (PAGE) gels of material from the supernatants of cultures grown under shaking conditions (9).
Previously, we tried to increase pilus expression by laboratory-grown serovar Typhi. Insertion of a tac promoter upstream of pil did not enhance pilus production, possibly because overproduction of PilL (an outer membrane protein [14]) and/or PilM (an inner membrane protein [14]) destabilized bacterial membrane function. We then placed the pilN-pilV genes of serovar Typhi J341 under tac promoter control (21) while we allowed the proximal pil genes pilL and pilM to remain under pil promoter control. We constructed mutants with alterations in the pilS, pilV, or rci genes (9). In the wild-type strain an active Rci gene product may invert shufflon DNA to place either of two distinct C termini on a constant N-terminal pilV region, thus creating two distinct pilV genes designated pilV1 and pilV2. The pilV mutant does not express any PilV protein. Mutants in which rci was inactivated expressed either the pilV1 or pilV2 gene constitutively. A pilS mutation has been described previously (21). Notably, all strains except the pilS control strain expressed PilS protein in the supernatants of cultures grown (with shaking) in a manner that permitted shearing of pili (9). While a pilV mutant and a wild-type strain (both lacking the tac promoter) self-associated and a pilS mutant did not self-associate, it was noted that pilV1(Con) and pilV2(Con) mutants (lacking the tac promoter; formerly published as pilV1C and pilV2C) were also incapable of self-association, even though such strains seemed perfectly able to synthesize pili. To explore this apparent anomaly, the serovar Typhi strains with the pilN-pilV genes under tac promoter control were grown in static cultures for 48 h and examined by electron microscopy (Fig. 1). A strain with a wild-type shufflon region was heavily piliated (Fig. 1A), as was the pilV mutant (Fig. 1B). We prepared grids from each of five different cultures of both of the strains with the wild-type shufflon and the strain with the pilV mutation and examined 100 bacteria on each grid. We found that 53% of the wild-type bacteria were piliated, as were 90% of the organisms harboring the pilV mutation. This suggested that PilV expression inhibited pilus production. We measured the lengths of the pili on each of 100 bacteria on each grid and calculated that 14% of wild-type pili were >300 nm long, while the corresponding value for pili synthesized by the pilV mutant was 33%. The wild-type pili were therefore shorter on average than the pili synthesized by the pilV mutant. As expected, a pilS mutant lacked pili (Fig. 1C). Notably, the strains bearing either the pilV1(Con) or pilV2(Con) mutation were also nonpiliated (Fig. 1D and E). We used immunogold labeling to show that the pili of the strains with either the wild-type shufflon or the pilV gene were assembled from PilS subunits (Fig. 2A and B). We detected pilus tip PilV protein by immunogold labeling (Fig. 2C) on only a minority (<1%) of wild-type pili. We discuss this finding below. Exposure of grids to preimmune sera did not permit subsequent gold labeling of pilus components by the secondary antibody.
It is important to ask if strains bearing the tac promoter before pilN produce normally assembled pili. The micrographs in Fig. 1A and 1B show the production of recognizable pili under such conditions. The tac insertion should result in strong transcription enhancement of (among other pil genes) both pilS and pilV. There is no reason to suggest that the ratio of pilS transcription to pilV transcription in the strains used here is different from the ratio in a wild-type strain. When strains with normal pil operons were used in self-association tests (9), it was clear that PilV expression eliminated self-association under conditions under which self-association was measurable with wild-type bacteria or a pilV mutant. This may be explained if PilV production eliminates pilus assembly in wild-type serovar Typhi, as was found in the experiments using tac-driven pilN-pilV described here.
As the strains bearing either the pilV1(Con) or pilV2(Con) mutation were nonpiliated but expressed apparently normal (wild-type) amounts of PilS protein in the supernatants of cells grown in shaking cultures (9), it seemed possible that constitutive expression of a PilV protein resulted in pilus detachment even when cells were grown in the absence of shearing forces. We wished to minimize growth conditions that might shear pili, but growth in static cultures was very slow; the stationary phase was entered by 1-liter cultures in 2.8-liter flasks only after ca. 72 h of incubation at 37°C. We found that shaking growth conditions (i) permitted the strains to reach the stationary phase after ca. 20 h of growth but (ii) did not seem to cause pilus shearing provided that the depth of the culture medium was such that simple swirling without obvious agitation of the culture was observed. We therefore prepared 1-liter cultures in 2-liter flasks. The medium depth was 6.2 cm in these flasks. In our previous work (9) 250-ml cultures were grown in 1-liter flasks with a culture depth of 2.2 cm and associated agitation during shaking growth. Small volumes (20 ml) of stationary-phase cultures were inoculated into 1 liter of fresh LB medium and shaken (220 rpm) at 37°C for 20 h. The optical densities of the cultures were standardized, and the cultures were centrifuged at 5,000 x g for 20 min at 4°C. The supernatants were then centrifuged at 13,000 x g for 10 min at 4°C to remove all bacterial cells. The supernatants from this step were spun for 1 h in a Sorvall A-641 fixed-angle ultracentrifuge rotor operating at 186,453 x g and 4°C. The pelleted material from 1 liter of supernatant was resuspended in 600 μl of phosphate-buffered saline (pH 7.4), and this preparation was designated the crude pilus fraction.
To prepare a linear sucrose gradient, sucrose solutions containing 20, 30, 40, 50, and 60% (wt/wt) sucrose in phosphate-buffered saline (pH 7.4) were sequentially layered (2 ml each) to form a 10-ml step gradient, which was then allowed to diffuse for 14 h at 4°C. The crude pilus fraction (from ca. 1012 bacterial cells) was loaded and subjected to 45 min of centrifugation in a Sorvall TH-641 swinging-bucket rotor at 106,394 x g and 19°C. Fourteen fractions that were the same volume were collected from the top of the gradient. Aliquots (15 μl) of each fraction were analyzed by SDS-PAGE, followed by Western blotting using either mouse anti-prePilS (1:10,000) (11) or mouse anti-His6-prePilV1 (1:5,000) as the primary antiserum (Fig. 3). The supernatants of strains either with the wild-type shufflon (Fig. 3A) or with the pilV gene (Fig. 3B) contained very few pili. Traces of PilS protein were found in fractions 6 to 11 of the gradients, which presumably reflected minor shearing of full-length pili during the initial low-speed centrifugation step. A small amount of PilV protein was found in the culture supernatant of the strain with the wild-type shufflon (Fig. 3C). The PilV protein did not appear to be preferentially associated with the pili separated on the same gradient but rather remained in the top fraction, perhaps reflecting facile removal of PilV proteins from pilus tips, as discussed below. Some PilV protein was also found in the bottom gradient fractions. This may have represented PilV protein in bacterial membrane fragments in which the protein may have resided for a time prior to secretion. No PilV protein was found in culture supernatant from the strain carrying the pilV mutation (Fig. 3D). As expected, the pilS mutant strain expressed neither the PilS (Fig. 3E) nor PilV (Fig. 3F) protein. The pilS mutation, which resulted from insertion of a Kmr cassette into the pilS gene, should be polar on expression of the downstream pilV gene.
Notably, culture supernatants of strains carrying either the pilV1(Con) or pilV2(Con) mutation were rich in both the PilS (Fig. 3G and H) and PilV (Fig. 3J and K) proteins. In either case, the PilS proteins were distributed fairly evenly across gradient fractions, showing that pili of various sizes were first synthesized by cells of these strains and then detached, even when the cells were grown under conditions that minimized shearing of pili. As the PilV1 and PilV2 proteins are constitutively expressed in these strains, it was not surprising (given the secretion of PilV proteins by the strain with the wild-type shufflon, as mentioned above) to find that PilV proteins were prominent in culture supernatants. As previously noted in the case of PilV expression by the strain with the wild-type shufflon, the PilV1 and PilV2 proteins seemed to be enriched (i) in the top (less dense) gradient fractions and (ii) in the bottom (more dense) gradient fractions, where sizable membrane fragments containing PilV proteins which had not yet been exported would be expected. With the pilV1(Con) strain, both prePilV1 and PilV1 were found in the membrane-enriched (bottom) fractions, as double bands reactive with the anti-His6-prePilV antiserum appeared.
It is very likely that the PilV proteins are normally located at pilus tips because strains bearing various possible R64-encoded PilV proteins discriminate between bacterial recipients in liquid mating preparations (3, 20). The PilV proteins recognize determinants in the lipopolysaccharide molecules of the recipients, thus establishing connections with the donor (the R64-bearing strain) to facilitate transfer of the R64 plasmid. The PilV proteins do not seem to be stably pilus associated, however. With the strain expressing wild-type shufflon, however, the PilV proteins were not in the pilus-containing fractions. It is possible that the manipulations (high-speed centrifugation and pellet resuspension) required for preparation of material for sucrose gradient velocity centrifugation tend to result in detachment of PilV proteins from pilus tips. If this is true, then the favored localization of secreted PilV protein in the first gradient fraction [in both wild-type and pilV1(Con) and pilV2(Con) mutant strains] is explained. With the pilV1(Con) and pilV2(Con) mutant strains, smaller amounts of PilV proteins may be dispersed through the gradients. It seems likely that the high-frequency detachment of pili in strains constitutively expressing PilV proteins leads to high levels of PilV-tipped pili in the culture supernatants, and not all the PilV protein molecules are detached from the pili during manipulations prior to sucrose gradient velocity sedimentation. The suggested ease of detachment of PilV proteins from pilus tips may explain why it was difficult to find PilV-tipped pili by electron microscopy (Fig. 2C). While some long pili separated on the sucrose gradient may have retained pilus tip PilV proteins (Fig. 3J and K), immunogold microscopy of sucrose gradient fractions did not consistently reveal pilus tip PilV molecules, as manipulations prior to electron microscopy probably removed the already small amounts of pilus tip PilV proteins that remained pilus associated during sucrose gradient sedimentation. The data do not permit rigorous exclusion of the possibility that PilV secreted independently into culture medium acts externally to fracture pili containing exclusively PilS protein. If this occurred, however, then all fracture points would be at the outer membrane surface in pilV1(Con) and pilV2(Con) mutant strains (Fig. 1D and E) but remote from the membrane surface in the wild-type strain (Fig. 1A). We consider this unlikely.
The data described above allow us to propose that incorporation of a PilV protein into pilus growth at a site previously processing only PilS molecules causes detachment of the preexisting pilus. In serovar Typhi, pilus growth need not commence with a PilV protein, as the strain with the pilV gene is heavily piliated. In the strain with the wild-type shufflon a small amount of PilV protein appears in culture supernatants, compared to the levels seen in culture supernatants from strains constitutively expressing PilV1 and PilV2 proteins (Fig. 3, compare panel C with panels J and K). As we have previously shown, PilV protein expression is inhibited when shufflon inversion inhibits through-transcription of the pilV gene (16). The average pilus length of the strain with a wild-type shufflon was less than average pilus length of the pilV mutant (see above). This suggests that a small amount of PilV protein synthesis by the wild-type strain may indeed be involved in some pilus detachment. There were somewhat more pili in the culture supernatant of the strain with the wild-type shufflon compared with the very small amount of pili found in the culture supernatant of the strain carrying the pilV mutation (Fig. 3, compare panels A and B).
The serovar Typhi (and serovar Dublin and Paratyphi C [10, 15]) pil operons are clearly derived from a pil operon that is similar or identical to operons found on the enterobacterial plasmids R64 and R721. The shufflons of these plasmids are more complicated than that of serovar Typhi, and they have five or six 19-bp sequences, respectively, which serve as substrates for the rci-encoded recombinase (1, 2, 5). There is no evidence that serovar Typhi is a donor of DNA. Instead, we have suggested previously that the reduction in shufflon complexity in serovar Typhi to a simple two-state arrangement exploits the ability of PilV proteins to control pilus expression (9).
ACKNOWLEDGMENTS
This work was supported by the Hong Kong Government Research Grants Council.
We thank Tze Kin Cheung of the Materials Characterization and Preparation Facility of our university for his assistance with electron microscopy.
FOOTNOTES
Corresponding author. Present address: P.O. Box 20, Redlynch, Cairns, QLD 4870, Australia. Phone: (617) 4039-0939. E-mail: jhackett@cairns.net.au.
Present address: School of Optometry, University of California, Berkeley, CA 94720-2020.
Contact Christina Morris for strains or biological materials. E-mail: bctina@ust.hk.
REFERENCES
1. Gyohda, A., and T. Komano. 2000. Purification and characterization of the R64 shufflon-specific recombinase. J. Bacteriol. 182:2787-2792.
2. Gyohda, A., N. Furuya, N. Kogure, and T. Komano. 2002. Sequence-specific and non-specific binding of the Rci protein to the asymmetric recombination sites of the R64 shufflon. J. Mol. Biol. 318:975-983.
3. Ishiwa, A., and T. Komano. 2003. The pilus PilV adhesins of plasmid R64 recognize specific structures of the lipopolysaccharide molecules of recipient cells. J. Bacteriol. 185:5192-5199.
4. Ivanoff, B. 1995. Typhoid fever: global situation and WHO recommendations. Southeast Asian J. Trop. Med. Public Health 26:1-6.
5. Kim, S.-R., and T. Komano. 1992. Nucleotide sequence of the R721 shufflon. J. Bacteriol. 174:7053-7058.
6. Komano, T., S.-R. Kim, T. Yoshida, and T. Nisioka. 1994. DNA rearrangement of the shufflon determines recipient specificity in liquid mating of IncI1 plasmid R64. J. Mol. Biol. 243:6-9.
7. Kubo, A., A. Kusukawa, and T. Komano. 1988. Nucleotide sequence of the rci gene encoding shufflon-specific DNA recombinase in the IncI1 plasmid R64: homology to the site-specific recombinases of integrase family. Mol. Gen. Genet. 213:30-35.
8. Lyczak, J. B., and G. B. Pier. 2002. Salmonella enterica serovar Typhi modulates cell surface expression of its receptor, the cystic fibrosis transmembrane conductance regulator, on the intestinal epithelium. Infect. Immun. 70:6416-6423.
9. Morris, C., I. S. M. Tsui, C. M. C. Yip, D. K.-H. Wong, and J. Hackett. 2003. The shufflon of Salmonella enterica serovar Typhi regulates type IVB pilus-mediated bacterial self-association. Infect. Immun. 71:1141-1146.
10. Morris, C., C. K. P. Tam, T. S. P. Wallis, P. W. Jones, and J. Hackett. 2003. Salmonella enterica serovar Dublin strains which are Vi antigen-positive use type IVB pili for bacterial self-association and intestinal cell entry. Microb. Pathog. 35:279-284.
11. Pan, Q., X.-L. Zhang, H.-Y. Wu, P.-W. He, F. Wang, M.-S. Zhang, J.-M. Hu, B. Xia, and J. Wu. 2005. Aptamers that preferentially bind type IVB pili and inhibit human monocytic-cell invasion by Salmonella enterica serovar Typhi. Antimicrob. Agents Chemother. 49:4052-4060.
12. Parry, C. M., T. T. Hien, G. Dougan, N. J. White, and J. J. Farrar. 2002. Typhoid fever. N. Engl. J. Med. 347:1770-1782.
13. Pier, G. B., M. Grout, T. Zaidi, G. Meluleni, S. S. Mueschenborn, G. Banting, R. Ratcliff, M. J. Evans, and W. H. Colledge. 1998. Salmonella typhi uses CFTR to enter intestinal epithelial cells. Nature 393:79-82.
14. Sakai, D., and T. Komano. 2002. Genes required for plasmid R64 thin-pilus biogenesis: identification and localization of products of the pilK, pilM, pilP, pilR, and pilT genes. J. Bacteriol. 184:444-451.
15. Tam, C. K. P., J. Hackett, and C. Morris. 2004. Salmonella enterica serovar Paratyphi C carries an inactive shufflon. Infect. Immun. 72:22-28.
16. Tam, C. K. P., J. Hackett, and C. Morris. 2005. Rate of inversion of the Salmonella enterica shufflon regulates expression of invertible DNA. Infect. Immun. 73:5568-5577.
17. Tsui, I. S. M., C. M. C. Yip, J. Hackett, and C. Morris. 2003. The type IVB pili of Salmonella enterica serovar Typhi bind to the cystic fibrosis transmembrane conductance regulator. Infect. Immun. 71:6049-6050.
18. Wang, F., X.-L. Zhang, Y. Zhou, L. Ye, Z. Qi, and J. Wu. 2005. Type IVB piliated Salmonella typhi enhance IL-6 and NF-B production in human monocytic THP-1 cells through activation of protein kinase C. Immunobiology 210:283-293.
19. Wu, H.-Y., X.-L. Zhang, Q. Pan, and J. Wu. 2005. Functional selection of a type IV pilus-binding peptide that specifically inhibits Salmonella Typhi adhesion to/invasion of human monocytic cells. Peptides 26:2057-2063.
20. Yoshida, T., N. Furuya, M. Ishikura, T. Isobe, K. Haino-Fukushima, T. Ogawa, and T. Komano. 1998. Purification and characterization of thin pili of IncI1 plasmids ColIb-P9 and R64: formation of PilV-specific aggregates by type IV pili. J. Bacteriol. 180:2842-2848.
21. Zhang, X.-L., I. S. M. Tsui, C. M. C. Yip, A. W. Y. Fung, D. K.-H. Wong, X. Dai, Y. Yang, J. Hackett, and C. Morris. 2000. Salmonella enterica serovar Typhi uses type IVB pili to enter human intestinal epithelial cells. Infect. Immun. 68:3067-3073.(Connie K. P. Tam, Christina Morris, and )