David W. Holden. E-mail firstname.lastname@example.org; Tel. (+44) 181 383 3487; Fax (+44) 181 383 2078.
The type III secretion system of Salmonella pathogenicity island 2 (SPI-2) is required for bacterial replication inside macrophages. SseB has been considered a putative target of the secretion system on the basis of its similarity with EspA, a protein secreted by the type III secretion system of enteropathogenic Escherichia coli (EPEC). EspA forms a filamentous structure on the bacterial cell surface and is involved in translocation of proteins into the eukaryotic cytosol. In this paper, we show that SseB is a secreted protein that associates with the surface of the bacterial cell and might, therefore, also be required for delivery of SPI-2 effector proteins to the eukaryotic cell cytosol. SseB begins to accumulate inside the bacterial cell when the culture enters early stationary phase. However, SseB is only secreted if the bacteria are grown at low pH or if the pH is shifted after growth from 7.0 to below pH 5.0. The secretion occurs within minutes of acidification and is totally dependent on a functional SPI-2 type III secretion system. As the pH of the Salmonella-containing vacuole inside host cells has been shown to acidify to between pH 4.0 and 5.0, and as SPI-2 gene expression occurs inside host cells, low pH might be a physiological stimulus for SPI-2-mediated secretion in vivo.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Type III secretion systems of pathogenic bacteria comprise a large number of proteins that transfer specific effector proteins from bacterial into eukaryotic cells. This process typically involves a secreton for exporting proteins from the bacterial cell, a translocon for transferring effector proteins into host cells, various regulators that control gene transcription and protein secretion, chaperones that bind to and are needed for secretion of translocators and effector proteins and the effector proteins themselves (Hueck, 1998). Effector proteins can modulate host cell signalling pathways in a variety of ways leading, for example, to bacterial uptake in the case of Shigella (Adam et al., 1995; Ménard et al., 1996) or the blocking of phagocytosis in the case of Yersinia (Cornelis and Wolf-Watz, 1997).
Salmonella typhimurium is unusual in that it contains two type III secretion systems. One, called Inv/Spa, controls bacterial invasion of epithelial cells (Galán and Curtiss, 1989; Galán, 1996). Most of the genes associated with Inv/Spa are encoded within a 40 kb pathogenicity island (SPI-1) located at 63 centisomes (cs) on the chromosome (Mills et al., 1995). Genes for the second type III secretion system are found within a second pathogenicity island (SPI-2) at 30 cs on the chromosome. This secretion system plays a crucial role in systemic growth of Salmonella in its host (Hensel et al., 1995; Shea et al., 1996) and is required for bacterial growth in macrophages (Ochman et al., 1996; Cirillo et al., 1998; Hensel et al., 1998).
The identification and characterization of type III-secreted proteins of Yersinia spp., Shigella spp., pathogenic Escherichia coli and Inv/Spa of Salmonella spp. has been made possible by the discovery of in vitro growth conditions in which the secreted proteins are exported (reviewed by Hueck, 1998). Although it has been demonstrated that SPI-2 genes are expressed inside host cells (Valdivia and Falkow, 1997) and can be regulated by certain environmental conditions, including phosphate starvation and low Mg2+ concentrations (Deiwick et al., 1999), no secreted protein of the SPI-2 system has yet been identified, and the conditions that trigger protein secretion by this system have not been defined.
On the basis of similarities of sequence and organization to genes of other bacterial pathogens, several genes located in a 9 kb region within SPI-2 have been proposed to encode secreted proteins (Hensel et al., 1998). Proteins encoded by sseB, sseC and sseD display weak similarity to EspA, EspD and EspB respectively. These Esp proteins are secreted by the type III secretion system of the locus of enterocyte effacement (LEE) of enteropathogenic E. coli (EPEC) and Shiga toxin-producing E. coli (STEC) (Donnenberg et al., 1993; Kenny et al., 1996; Lai et al., 1997; Ebel et al., 1998) and are essential for virulence (Frankel et al., 1998). Recently, it has been demonstrated that EspA is involved in the translocation of EspB (Knutton et al., 1998; Wolff et al., 1998) and another LEE-encoded protein, Tir (Kenny et al., 1997), into the plasma membrane and cytoplasm of epithelial cells. EspA is a component of a filamentous surface appendage with a cylindrical structure that forms a contact between the bacterium and the eukaryotic cell surface (Ebel et al., 1998; Knutton et al., 1998) and could act as a channel through which proteins are delivered into the host cells (Frankel et al., 1998).
An S. typhimurium strain carrying a deletion of sseB is strongly attenuated in mice and is defective in replication within macrophages (Hensel et al., 1998). In view of this observation and the similarity between the sseB gene product and EspA, SseB seemed a likely candidate for a protein that might function as part of a SPI-2 translocon and could be used to establish in vitro conditions inducing SPI-2-mediated protein secretion. Using antibodies raised against recombinant SseB, we show here that SseB is rapidly secreted onto the bacterial cell surface by the SPI-2 type III secretion system when the pH of the culture medium is shifted to 5.0. This pH-based induction may be physiologically relevant, as the pH of the Salmonella-containing vacuole of infected cells has been reported to drop to 4.0–5.0 within 1 h after bacterial uptake (Rathman et al., 1996). Furthermore, as the Inv/Spa system is induced at pH 8.0 (Daefler, 1999), it appears that Salmonella regulates the two secretion systems independently in a way that prevents the potentially harmful consequences of activating two structurally related, but functionally independent, secretion systems in the same cell.
Surface localization of SseB
To investigate SseB secretion, polyclonal antiserum was raised against recombinant SseB, and this was used to detect SseB by immunoblot analysis. Initial experiments showed that the intracellular levels of SseB increased when bacteria were grown in low-Mg2+ or low-phosphate media (data not shown), as has been reported for SsaP and SscA (Deiwick et al., 1999). However, under these conditions, no secretion of SseB was detected using the methods described below (data not shown). As SPI-2 gene expression is induced inside host cells (Valdivia and Falkow, 1997), we grew S. typhimurium wild-type strain 12023 at 37°C overnight in intracellular salt medium (ISM), which reflects the salt conditions inside eukaryotic cells (Headley and Payne, 1990). The pH of the medium was adjusted to 5.0, as the pH of the Salmonella-containing vacuole decreases to between 4.0 and 5.0 shortly after bacterial entry (Rathman et al., 1996).
To determine whether SseB is an intracellular protein, a protein secreted to the cell surface or is secreted into the culture medium, we treated bacterial cultures to obtain three fractions. The supernatant fraction contains proteins recovered by precipitation of the culture supernatant after centrifugation to isolate the bacterial cells. The second fraction consists of proteins detached from the surface of pelleted cells by the hydrophobic agents p-xylene or n-hexadecane (Michiels et al., 1990; Ménard et al., 1994; Iriarte et al., 1998). These agents allow efficient recovery of proteins weakly associated with the bacterial cell surface, but not of known membrane proteins (Michiels et al., 1990). The cell pellet fraction contains proteins remaining in the cell pellet after these treatments and includes intracellular proteins.
When proteins in the three fractions obtained from a culture of strain 12023 after growth in ISM pH 5.0 were analysed by immunoblotting for the presence of SseB, a polypeptide of the size predicted for SseB was detected in the xylene and cell pellet fractions, but not in the supernatant fraction, suggesting that SseB is mainly secreted onto the bacterial cell surface (Fig. 1A). Identical results were obtained using hexadecane as an alternative to xylene (data not shown). Constitutive expression of SseB from a plasmid resulted in an increased intensity of the signal in the xylene and cell pellet fractions but, again, it was not detected in the supernatant fraction (Fig. 1A).
To verify that the observed signals correspond to the SseB protein, we analysed fractions from cultures of strain 12023 and an isogenic ΔsseB strain by immunoblotting. A single polypeptide was detected in the pellet and hexadecane fractions from the wild-type strain, but not from the ΔsseB strain (Fig. 1B). The antibody specificity was also confirmed by immunoprecipitation experiments and by expression of SseB in E. coli DH5α (data not shown).
To determine whether hexadecane treatment had a significant effect on bacterial cell integrity, serial dilutions of a culture of strain 12023 were plated to LB medium before and after extraction with hexadecane. There was no decrease in the numbers of bacterial colony-forming units (cfu) after treatment, indicating that exposure to hexadecane did not adversely affect cell viability (data not shown).
We also found that SseB was removed from the bacterial cell surface by vigorous vortexing, by passing the culture through a 26-gauge needle and by treating cells with 1 M Tris-Cl, pH 7.5 (data not shown).
A recent report has shown that SipC and InvJ, two Inv/Spa-secreted proteins, could be recovered from the plastic surface of the vessel in which the bacterial cells were cultured (Daefler, 1999). When strain 12023 was grown under the conditions described above, a small amount of SseB was also found on the surface of the growth vessel (Fig. 1C). Therefore, detection of extracellular SseB does not require prior treatment of the bacterial cells. These results indicate that the extracellular location of SseB is not an artifact of the fractionation methods.
Approximately half the total amount of SseB was detected in the hexadecane fraction, but other means of detaching it from the bacterial cell surface were less efficient. Approximately 10% of total SseB was recovered from the plastic surface of the growth vessel after overnight growth. Therefore, in subsequent experiments, hexadecane was routinely used to assay secreted SseB.
As a further test for the presence of SseB on the cell surface, bacterial cells grown overnight under inducing conditions were treated with proteinase K. Proteinase K has been widely used to discriminate between intracellular and extracellular proteins, because its addition to the bacterial cell culture results in degradation of extracellular but not intracellular proteins (Ménard et al., 1994; Iriarte et al., 1998). After proteinase K incubation, the enzyme was inactivated, and cells were treated with hexadecane. In this case, SseB was no longer recovered, and the amount of SseB detected in the cell pellet fraction was not affected (Fig. 2A). Hence, the hydrophobic agents, xylene and hexadecane, can efficiently separate surface-localized SseB from that in the remainder of the cell.
In contrast to SseB, the SPI-2-encoded proteins SscA and SsaP were detected in the cell pellet but not in the hexadecane fractions (Fig. 2B). SscA is a putative chaperone (Hensel et al., 1998), and SsaP is similar to SpaN, an outer membrane protein of the type III secretion system of Shigella flexneri, and to InvJ, a Inv/Spa-secreted protein (Hueck, 1998).
Collectively, these data show that secretion of SseB can be assayed by hexadecane fractionation, that this secretion is induced during growth in laboratory medium and that the majority of the secreted protein remains weakly bound to the bacterial cell surface.
Effect of different mutations on SseB secretion
To establish whether SseB secretion was dependent on the type III secretion system encoded by SPI-2, several strains carrying mutations in genes encoding putative secretion apparatus proteins were analysed for SseB secretion. SsaV is predicted to be an inner membrane protein (Hensel et al., 1997) and shares strong similarity with the LcrD family of proteins (Galán et al., 1992). SsaJ is a predicted lipoprotein (Hensel et al., 1997) that may link inner and outer membranes based on its similarity with the YscJ/Mxi family of lipoproteins (Michiels et al., 1991). SsaC is similar to InvG (Ochman et al., 1996; Shea et al., 1996), a member of the PulD family of translocases (Kaniga et al., 1994) and an outer membrane protein (Crago and Koronakis, 1998). Mutations in the genes encoding these proteins prevented secretion of SseB, because no SseB was detected on the bacterial cell surface or on the surface of the growth vessel (Figs 1C, 3A and data not shown). However, these mutations did not significantly affect the intracellular pool of SseB (Figs 1C and 3A). The ssrA gene is predicted to encode the sensor protein of a two-component regulatory system (Shea et al., 1996) and is required for efficient transcription of other SPI-2 genes (Hensel et al., 1998). As expected, in the ssrA mutant, SseB was not detected in either fraction (Fig. 3A).
To confirm that SseB secretion under inducing conditions was occurring solely through the SPI-2 secreton, we analysed its secretion in a strain carrying a mutation in prgH, which encodes a structural protein of the Inv/Spa apparatus (Behlau and Miller, 1993), and a strain carrying a mutation in hilA, a gene encoding a regulator of the inv/spa genes (Bajaj et al., 1995). The levels of secreted SseB in these strains were similar to that of the wild-type strain (Fig. 3A).
To determine if other SPI-2 genes might play a role in SseB expression or secretion, we analysed SseB location in several other mutant strains. Non-polar mutations in sseC, sseE, sseF and sseG had no effect on levels of SseB in the hexadecane or pellet fractions (Fig. 3B). SseC has sequence similarity to a class of proteins involved in the translocation of effectors into the target host cell and contains three hydrophobic regions that could represent membrane-spanning domains (Hensel et al., 1998). SseE, SseF and SseG have no significant similarities to other known proteins. The three proteins have predicted transmembrane helices varying between one for SseE to four for SseF (Hensel et al., 1998). These proteins might represent part of a translocon and not be involved with SseB secretion onto the bacterial cell surface, or they might be effectors that are delivered into the host cell.
SseB secretion is induced by acid pH
Different media were tested for SseB secretion in an attempt to identify the component(s) responsible for the induction of secretion. Growth in media with either low phosphate (ISM-MES) or low Mg2+ (MgM medium) increased the total amount of SseB detected at a pH of either 5.0 or 7.5, compared with growth in ISM (Fig. 4). However, in all the media tested, SseB was only recovered from the cell surface when the bacteria were cultured at pH 5.0. When strain 12023 was grown in MgM-MES medium, SseB was detected on the bacterial cell surface and on the growth vessel at pH 5.0, but not at pH 7.5. There was no significant difference in the levels of intracellular SseB at either pH (Fig. 1C).
To investigate whether this apparent acid-induced secretion merely reflects an increase in sseB transcription, the SseB protein was expressed from a plasmid under the control of a constitutive promoter in MgM medium at pH 5.0 and pH 7.5. Secretion of SseB was again only detected at pH 5.0 (Fig. 5A), indicating that the pH-induced secretion of SseB appears to be independent of sseB transcription. In addition, transcription of sseB was analysed using a transcriptional fusion to sseA, which lies immediately upstream and is part of the same operon as sseB (Cirillo et al., 1998). The sseA::luc fusion strain was grown in MgM medium at pH 7.0 until the OD600 reached ≈0.5. The medium was then changed to fresh MgM medium at pH 5.0 or pH 7.0, and luciferase activity was determined at different time points. Exposure to the lower pH did not increase transcription of the gene fusion (Fig. 5B). We conclude that the pH-based control of SseB secretion is not caused by transcriptional activation. Rather, these observations suggest that SseB delivery to the surface relies on the acidic pH-induced release of preformed molecules of SseB stored in the cytosol.
Growth phase regulation of SseB expression
To investigate the kinetics of the pH-induced secretion, strain 12023 was grown in MgM medium, pH 7.5, at 37°C and, when the OD600 of the culture reached ≈0.35, the medium was replaced with fresh MgM medium at pH 5.0, and bacteria were sampled at different time points of growth. Over the first hour, SseB was barely detected in the cell pellet fraction of the culture and was not detected in the hexadecane fraction (Fig. 6A). When the bacteria reached an OD600 of ≈0.9 (entering stationary phase), the amount of SseB detected inside the cells increased greatly, and SseB was detected in the hexadecane fraction (Fig. 6A). A mock induction experiment was performed in the same way, except that the pH was kept at 7.5. The same kinetics of intracellular accumulation of SseB were observed, but no secreted SseB was detected (Fig. 6B).
To rule out the possibility that the appearance of SseB in the hexadecane fraction resulted from the release of intracellular SseB from lysed cells in the bacterial culture, the same experiment was carried out using a strain carrying a mutation in ssaV, encoding one of the components of the secretion apparatus. A similar accumulation of intracellular SseB was observed, but no secreted protein was detected (Fig. 6C).
Rapid secretion of SseB
To investigate the kinetics of SseB secretion in closer detail, strain 12023 was grown in MgM medium, pH 7.0, at 37°C until the cells were entering stationary phase. At this time, the pH of the medium was shifted to 5.0, and bacteria were sampled at different time points thereafter. Measurement of the culture medium pH confirmed that it did not change during the course of the experiment. After 15 min of incubation at pH 5.0, SseB was detected in the hexadecane fraction. There was no obvious change in the amount of intracellular SseB after the pH shift, confirming that a drop in pH does not dramatically affect SseB production. A similar amount of SseB was recovered from the cell surface after 1 h of incubation at acidic pH (Fig. 7A), suggesting that pH induction results in a quick release of a proportion of presynthesized SseB.
To rule out the possibility that pH microchanges might affect SseB secretion, we carried out an induction experiment by transferring bacteria to fresh MgM-MES medium, pH 5.0, a medium with good buffering capacity in this range. Strain 12023 was grown in MgM medium, pH 7.0, until the cells were entering stationary phase, and cell aliquots were transferred to MgM-MES medium, pH 5.0. Bacteria were then sampled at different time points. Two minutes after the pH change, SseB was already detected in the hexadecane fraction, and longer incubation times did not result in greater yields (Fig. 7B). This yield represented ≈10% of the total SseB detected, significantly lower than the amount of secreted protein detected after overnight growth. Transfer of a cell aliquot to pH 7.0 medium did not lead to secretion of SseB (data not shown). The short time in which the secretion of SseB takes place after the pH shift strongly suggests that the secretion apparatus is assembled in the membrane before the pH change.
Effect of different external pH levels on SseB secretion
To characterize acid-induced secretion of SseB further, we analysed the SseB secretion response to different pH levels ranging from 7.5 to 4.5. Secretion of SseB was detected when strain 12023 was shifted to MgM medium at pH 4.5 or pH 5.0 after growth at pH 7.5, but not when it was transferred to medium at a higher pH (Fig. 8A). To confirm that the appearance of SseB in the hexadecane fraction was not the result of non-specific release caused by the acid environment, we included a strain carrying a mutation in ssaV as a control. No SseB was detected in the hexadecane fraction of this strain after exposure to pH 4.5, confirming that the secretion of SseB in the wild-type strain was mediated by SPI-2 (Fig. 8B).
Agglutination of bacterial cells
Attempts to identify structures on the surface of strain 12023 similar to those containing EspA in EPEC (Knutton et al., 1998) after growth in inducing conditions were not successful. However, a proportion of cells in such cultures were agglutinated by the anti-SseB serum (Fig. 9), indicating that SseB is exposed on the surface of the bacterial cell. It is possible that the SPI-2 secretion system does not produce filamentous structures containing SseB or, if it does, these appendages might be relatively fragile.
In this paper, we show that SseB is a component of the SPI-2 type III secretion system of S. typhimurium and is secreted onto the bacterial cell surface in response to acidification of the culture medium. Initial evidence implicating SseB as a protein secreted by the SPI-2 system was based on amino acid sequence similarity to EspA of EPEC. Further evidence comes from the similar phenotype of ΔsseB and other SPI-2 mutant strains in macrophages and in vivo, and the fact that transcription of the sseB operon is dependent on ssrA (Cirillo et al., 1998; Hensel et al., 1998). Expression of several other SPI-2 genes, including sseA, has been shown previously to be enhanced by low Mg2+ and phosphate starvation (Deiwick et al., 1999). As sseB is part of the same operon as sseA (Cirillo et al., 1998; T. Kubo and D. W. Holden, unpublished results), it is not surprising that greater amounts of SseB were found in cells growing in media with limiting Mg2+ or phosphate. However, like the other SPI-2 genes that have been examined (Deiwick et al., 1999), this increase occurred gradually and reached maximal levels after the bacterial cultures had entered stationary growth phase. These results suggest that, although nutrient deprivation is an important prerequisite for SPI-2 gene expression, there are other factor(s) that are present at stationary phase growth and are necessary for SseB accumulation. The identity of these factor(s) and the mechanism of this regulation are presently unknown.
Transfer of SseB to the bacterial cell surface was shown to be dependent on three genes predicted to encode components of the SPI-2 secreton, but mutations in SPI-1 genes did not affect this process. Therefore, SseB is the first secreted protein of the SPI-2 system to be identified. SseB secretion is apparently controlled by the pH of the medium, with maximal secretion occurring rapidly after transfer of cells to pH 4.5–5.0. The levels of secreted SseB in cultures grown overnight under inducing conditions were higher than in those induced by a pH shift. This suggests that the acid environment allows limited secretion and that the cells then enter a stage during which secretion is prevented. The higher levels of secreted SseB detected after overnight culture might result from several rounds of secretion. Acidification of the media did not have a significant effect on sseB expression or intracellular SseB levels, but appears to be the external stimulus that triggers the secretion of preformed cytosolic SseB.
Changes in extracellular pH influence production and secretion of components of other type III secretion systems. A decrease in extracellular pH from pH 7.4 to pH 6.0 resulted in a 10-fold decrease in virF transcription and reduced invasiveness of Shigella flexneri (Nakayama and Watanabe, 1995). High pH has been shown to increase the expression of type III secretion system genes in Pseudomonas syringae (Rahme et al., 1992; Xiao et al., 1992) and Erwinia amylovora (Wei et al., 1992). Recent work has shown that pH is probably the most important factor regulating the expression and secretion of components of the Inv/Spa system of S. typhimurium (Daefler, 1999). In this case, however, induction occurred after the pH of the media was shifted from acidic to mildly alkaline, and synthesis of machinery components and secretion of substrates were induced simultaneously after the pH change. Acid-induced secretion of SseB therefore appears to be a novel variation on the theme of pH regulation of type III secretion. Several type III secretion systems have been shown to be induced in vivo by intimate contact with the host cell membrane (Ménard et al., 1994; Rosqvist et al., 1994; Zierler and Galán, 1995; Petterson et al., 1996), leading to the concept that type III secretion is contact dependent (Galán, 1996; Cornelis and Wolf-Watz, 1997). The finding that acidic pH is sufficient to induce secretion of SseB by the SPI-2 system in laboratory conditions does not rule out a contact-dependent induction in vivo, and further work is required to establish the signals that regulate secretion inside host cells.
The InvA/Spa secretion system is involved in bacterial invasion of the gut epithelium of the small intestine (Galán, 1996), and the alkaline-controlled induction of SPI-1 genes is thought to reflect the increase in physiological pH as the bacteria pass from the stomach to the small intestine (Daefler, 1999). The SPI-2 secretion system is required for bacterial replication inside the macrophages (Ochman et al., 1996; Cirillo et al., 1998; Hensel et al., 1998), and SPI-2 genes are induced over several hours after bacterial entry into host cells (Cirillo et al., 1998). After uptake, the bacteria become localized in vacuoles that have been reported to acidify rapidly to pH 4.0–5.0 (Rathman et al., 1996). Intracellular bacterial accumulation and SPI-2 gene expression were reduced by inhibition of the acidification process by bafilomycin (Rathman et al., 1996; Cirillo et al., 1998). We and Cirillo et al. (1998) did not observe transcriptional induction of SPI-2 genes by low pH in vitro, suggesting that the effect of bafilomycin on SPI-2 transcription may not result from a direct effect on pH. These observations, together with the results presented here, suggest that low pH could be the physiological signal mediating SPI-2 secretion of SseB in host cells. Regulation of SPI-1 and SPI-2 secretion by different ranges of pH might help to ensure that these two functionally distinct systems are activated independently and only when and where they are required.
The ability of hydrophobic reagents to remove SseB from the cell indicates that the protein becomes weakly associated with the cell surface after secretion. EspA, which shares limited sequence similarity with SseB, is part of a filamentous organelle that links the bacterium and the eukaryotic cell and is required for the transfer of other secreted proteins into infected cells (Ebel et al., 1998; Knutton et al., 1998). Although initial attempts to identify SPI-2-dependent surface structures on S. typhimurium have not been successful, the results reported here support the notion that SseB could be part of the SPI-2 translocon delivering effector proteins from S. typhimurium into the host cell.
Bacterial strains and plasmids
The bacterial strains and plasmids used in this study are listed in Table 1. Strain HH129 was constructed by P22 transduction of a hilA ::Tn10 mutation from a derivative of SL1344 (a gift from Catherine Lee, Harvard Medical School, USA) to strain 12023. Strain HH130 was constructed by P22 transduction of a prgH ::TnphoA mutation from IB040, a derivative of SL1344 (Behlau and Miller, 1993), to strain 12023.
Intracellular salts medium (ISM) contained 170 mM K2PO4/KH2PO4, pH 5.0 or 7.5, 0.5 mM MgSO4, 1 μM CaCl2, 6 mM K2SO4, 5 mM NH4Cl, 5 mM NaCl, 0.4% glucose and 2 μg ml−1 nicotinic acid (Headley and Payne, 1990); ISM-MES media were identical except that 170 mM 2-[N-morpholino] ethane-sulphonic acid at pH 5.0 or 7.5 replaced the KPO4; magnesium minimal medium (MgM) contained 100 mM Tris-Cl, pH 5.0 or 7.4, 5 mM KCl, 7.5 mM (NH4)2SO4, 0.5 mM K2SO4, 1 mM KH2PO4, 8 μM MgCl2, 38 mM glycerol and 0.1% casamino acids (Hmiel et al., 1986); MgM-MES media were identical except that 170 mM 2-[N-morpholino]ethane-sulphonic acid at the corresponding pH replaced the Tris-Cl. Antibiotics kanamycin, ampicillin or chloramphenicol were added at 50 μg ml−1 when appropriate. Except where indicated, bacteria were grown at 37°C overnight with aeration.
To investigate SseB production and secretion, the strains were grown in MgM medium at pH 7.5 to two different stages of growth. To assay cells growing in exponential phase, the pH was changed when the OD600 was ≈0.35. To assay cells entering stationary phase, pH changes were carried out at an OD600 of ≈1.0. The pH changes were carried out in two different ways: by replacing the medium or by shifting its pH with a 0.1 M solution of HCl. In both cases, the pH of the culture was monitored during the incubation period, and the results obtained were similar. Media replacement was done both by centrifuging the culture at 35 000 g for 5 min at 4°C and gentle resuspension of the cell pellet in fresh media, and by filtering the cells onto 0.45 μm HA filters (Millipore), transferring the filter to the fresh media and gently resuspending the cells. Both methods gave similar results.
Preparation of cell fractions
Bacterial cell densities were determined by measurement of the OD600. To ensure that protein from equal numbers of cells was analysed, protein samples were adjusted to OD600 values in all experiments such that each fraction from a 10 ml culture of OD600 0.6 was taken up in 100 μl of protein denaturing buffer for gel electrophoresis.
Cell cultures were cooled on ice and centrifuged at 35 000 g for 5 min at 4°C, and the proteins present in the supernatant were collected by trichloroacetic acid precipitation. After removal of the culture medium, the cell pellet was washed with ice-cold PBS buffer and resuspended in 0.3 ml of the same buffer. The suspension was mixed gently with 0.2 ml of xylene or n-hexadecane for 5 min at room temperature and centrifuged at 10 000 g for 10 min at room temperature (Michiels et al., 1990). The organic layer was discarded, the aqueous layer mixed with 1.2 ml of acetone and held at −20°C to form a precipitate, and the cell pellet was stored at −70°C before analysis. The acetone precipitate was centrifuged, dried and also stored at − 70°C before analysis. All fractions were dissolved in protein-denaturing buffer before polyacrylamide gel electrophoresis.
To analyse the presence of SseB on the surface of the plastic growth vessel, and to allow a direct comparison with other fractions of the same sample, an appropriate amount of protein-denaturing buffer was added to the drained growth vessel and incubated for 60 min at 37°C.
For proteinase K treatment, cells were pelleted from a 50 ml culture, washed and suspended in 2.5 ml of PBS buffer. A 500 μl aliquot was held on ice, and 20 μl of 25 mM CaCl2 was added to second and third aliquots. Finally, 2.5 μl of 20 mg ml−1 proteinase K was added to the third aliquot, and both the second and the third aliquots were held at 37°C for 20 min, when 6 μl of 200 mM EGTA was added to both. All three aliquots were washed three times with cold PBS buffer. n-Hexadecane and cell pellet fractions were then prepared as described above.
Recombinant SseB protein and generation of antibodies
For the generation of recombinant SseB (rSseB), a 493 bp fragment was generated by polymerase chain reaction (PCR) using primers SseB-For (5′-CTCGGATCCCAGCAAAATCCGTTT-3′) and SseB-Rev (5′-TGAAAGCTTCATGAGTACGTTTTCTG-3′) introducing BamHI and HindIII sites, as indicated by the underlining. After digestion with BamHI and HindIII, the fragment was ligated to vector pQE30 (Qiagen). The resulting construct pJD13 in E. coli M15 [pREP] was used for the expression of rSseB, introducing an N-terminal His-tag fusion.
rSseB was expressed according to standard protocols (Qiagen) and purified by metal-chelating chromatography using ‘Hi-trap’ columns according to the instructions of the manufacturer (Pharmacia). The 19.3 kDa antigen was administered to rabbits by subcutaneous injection of purified protein (about 1.0 mg) emulsified with complete or incomplete Freund's adjuvant for initial or booster immunizations respectively (Harlow and Lane, 1988). Antibody titres were checked by Western blots using the purified antigen and total-cell fractions of E. coli expressing sseB under the control of the T7 promoter. The generation of anti-SscA and anti-SsaP sera has been described previously (Deiwick et al., 1999).
Polyacrylamide gel electrophoresis and Western analysis of proteins
Protein fractions were dissolved in the appropriate volume of protein-denaturing buffer containing 62.5 mM Tris-Cl, pH 6.8, 2% SDS, 5% β-mercaptoethanol, 10% glycerol and 0.02% bromophenol blue and held at 100°C for 5 min. Proteins were immediately separated on a 12% SDS–polyacrylamide gel using a tricine buffer system (Schägger and von Jagow, 1987). The gel was stained with Coomassie blue (Sambrook et al., 1989) or, for Western analysis, transferred to Immobilon-P membranes (Millipore) using a semi-dry blotting apparatus (Bio-Rad) with the buffer system of Kyhse-Andersen (1984). Western analysis used the ECL detection system under conditions recommended by the manufacturer (Amersham Life Science). Incubation of membranes with primary antibodies (rabbit anti-SseB, anti-SscA and anti-SsaP sera) was followed by incubation with anti-rabbit horseradish peroxidase conjugate (Amersham Life Science) as secondary antibody. Immunoreactive species were quantified as chemiluminescent signals on X-ray film by densitometry scanning.
Bacterial strains were grown with gentle aeration for 16 h in MgM media at pH 5.0 and pH 7.5. Aliquots (5 μl) of the cultures were mixed with the same volume of a 10-fold dilution of anti-SseB serum and incubated for 5 min at room temperature. The samples were then analysed by phase-contrast microscopy at magnification × 400.
†The first two authors contributed equally to this work.
This project was supported by grants from the MRC (UK) to David Holden, and DFG (Germany) (grant He. 1964/2-1) to Michael Hensel. C. R. Beuzón was supported by an EMBO fellowship. We wish to thank Thomas Nikolaus and Kate Unsworth for practical assistance.