INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Uropathogenic strains of Escherichia coli secrete a pore-forming cytotoxin called alpha-hemolysin, also known as HlyA (for reviews, see references 4 and 40). The hlyCABD genes mediate synthesis, activation, and secretion of alpha-hemolysin (20). hlyA encodes for a 110,000-molecular-weight inactive protoxin, which is posttranslationally modified into an active form by HlyC-mediated acylation at two lysine residues (15, 18, 34). While acylation is obligatory for HlyA's cytolytic activity, it is not essential for secretion into the medium (26) or for its ability to form pores in synthetic membranes (18). HlyA is exported through a type I secretory pathway in which proteins are exported across the bacterial envelope by an N-terminal signal sequence-independent manner (20, 30). Instead, HlyA's export signal lies at its C-terminal end (27).

HlyB and HlyD constitute the inner membrane transport complex for HlyA secretion. HlyB is a member of the ATP-binding cassette transport protein superfamily (12). Its large cytoplasmic C-terminal domain binds ATP, whose hydrolysis provides the primary energy source for HlyA secretion (37). HlyD has a short N-terminal cytoplasmic region, a single transmembrane domain and a large periplasmic domain (33). It is a member of the membrane fusion protein class of proteins that are found as a part of the type I secretory pathway in gram-negative bacteria (7). These proteins are proposed to bridge the inner and outer membranes to facilitate the transport of various substrates into the medium.

Wandersman and Delepelaire (38) first showed the requirement for TolC in alpha-hemolysin secretion. TolC is also involved in the secretion of unrelated toxins (14, 42). TolC is an outer membrane protein (23) that has been implicated in many diverse cellular functions. tolC null mutants display a pleiotropic phenotype, including colicin tolerance (25), hypersensitivity to detergents (8, 24, 41), chromosomal partitioning defect (13), and reduced OmpF expression (22, 24, 31). In vitro reconstitution experiments showed that TolC forms a peptide-specific channel (2).

The in vivo secretion of alpha-hemolysin is dependent on an intact lipopolysaccharide (LPS) core (1, 35). However, unlike HlyA, HlyB, and TolC mutants, which are defective in the secretion of toxin, the culture supernatant of these LPS mutants largely contain inactive alpha-hemolysin. It was proposed that an intact LPS core influences toxin's tertiary structure during or after its secretion into the medium. Mutations affecting LPS, and only indirectly TolC's biogenesis, have been shown to reduce alpha-hemolysin secretion (39).

A recent study carried out by Thanabalu et al. (37) showed that HlyB and HlyD form a translocase complex independent of HlyA or TolC. During secretion, HlyA interacts individually with HlyB, HlyD, and TolC; HlyA's interaction with TolC is observed only when the toxin molecule is properly engaged with the HlyBD translocase. HlyB's ATPase activity is not needed to form the translocation complex but for releasing toxin molecules from the complex into the medium. The recently solved crystal structure of TolC shows that it has a large -helical domain that extends 100 Å into the periplasmic space (17). The proximal end of TolC must interact with the accessory proteins to form a continuous conduit from the cytoplasm to the external surface for the transport of alpha-hemolysin into the medium.

In this study, we have taken a genetic approach to elucidate the role of TolC in alpha-hemolysin secretion. Since tolC null mutations confer a pleiotropic phenotype, missense mutations were sought that primarily affected the secretion of alpha-hemolysin without influencing other TolC-associated functions. A unique class of tolC mutants was obtained in which the steady-state secretion of alpha-hemolysin was unaffected but secreted toxin had significantly reduced cytolytic activity.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains, plasmids, and growth media. All bacterial strains and plasmids used in this study are listed in Table 1. The rich medium (Luria-Bertani [LB]) was prepared as previously described (34). The medium was supplemented with 25 µg of ampicillin/ml, 12.5 µg of chloramphenicol/ml, and 0.4 mM IPTG (isopropyl--D-thiogalactopyranoside) when needed. Blood agar plates were prepared by supplementing LB agar with 5% defibrinated sheep erythrocytes (Remel, Lenexa, Kans.).

Plasmid construction and mutagenesis. The tolC gene was cloned under the control of an IPTG-inducible promoter of pTrc99A (Pharmacia). For this two mutagenic primers were utilized: the forward primer (5'-GAATGCCCATGGGGAAATTGCTCCCCATTC-3') created a unique NcoI site (underlined), and the reverse primer (5'-GCGGCAGATAACCCGAAGCTTTACGGTTGCC-3') created a unique HindIII site (underlined). In creating the NcoI site, the second codon of the TolC signal sequence was changed from AAG (lysine) to GGG (glycine). tolC DNA was amplified from the chromosome by PCR, digested with NcoI and HindIII, and ligated into appropriately restricted pTrc99A. Under noninducing growth conditions (without IPTG), the level of plasmid-encoded TolC was extremely low, as determined by Western blot analysis, but in the presence of 0.4 mM IPTG its level was similar (85%) to that produced from the chromosomal gene.

SDS-PAGE and Western blot analyses. The French press lysis method was used to isolate cell envelopes from whole cells (21). Membrane proteins were subsequently analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as previously described (19). For Western blot analysis, protein samples were analyzed on mini (7 by 8 cm) SDS-polyacrylamide gels and transferred onto Immuno-Lite membranes (Bio-Rad) by using a Mini Trans-Blot electrophoretic transfer cell (Bio-Rad). Blots were incubated with primary antibodies raised against HlyA or TolC (1:5,000 dilution) for 2 h and then with the secondary antibody (goat--rabbit immunoglobulin G) for 1 h. Detection was carried according to the manufacturer's protocol (Bio-Rad).

Sensitivities to colicin E1, bacteriophage, and antibiotics. Strain K53 was grown to stationary phase and exposed to UV radiation for 90 s to induce colicin production. The UV-irradiated cells were diluted 1:50 and grown overnight. Cells were treated with chloroform and centrifuged at 2,000 × g for 10 min at room temperature in a GH-3.7 horizontal rotor (Beckman). Twofold dilutions were prepared from the supernatant, and 10 µl of each dilution was spotted over a lawn of bacterial strain that was to be tested for colicin E1 sensitivity. Sensitivity to bacteriophage was also done by a serial dilution method described. For antibiotic sensitivity, presoaked antibiotic disks (Difco) were placed on bacterial lawns prepared on LBA. The zones of inhibition (in millimeters) were measured after incubation of the plates for 18 h at 37°C.

Hemolysin enzyme assay. Hemolysin enzymatic assays were carried out as previously described (10). Cells were harvested at an optical density at 600 nm (OD₆₀₀) of 0.8 by centrifugation. The supernatant was filtered through 0.45-µm (pore-size) syringe filters (Acrodisc; Pall Corp.). To obtain hemolytic activity within a linear range, the filtered supernatant was diluted fivefold in an enzyme assay buffer composed of 20 mM CaCl₂, 10 mM Tris-HCl (pH 7.5), and 160 mM NaCl. Sheep erythrocytes were prepared for the assay by washing them three times in 0.9% NaCl. The enzyme assay reaction mixture (for each time point) contained 900 µl of enzyme assay buffer, 20 µl of washed erythrocytes, and 80 µl of the diluted culture supernatant. Enzymatic reactions were carried out at 37°C in a shaking water bath. Aliquots (800 µl) withdrawn at various time points were chilled immediately by placing them on ice. Unlysed erythrocytes were removed by centrifugation (5 min, 12,000 × g at 4°C), and 500 µl of the supernatant was used to measure the absorbance at OD₅₃₀ (as an indication of the amount of released hemoglobin). The hemolytic activity is defined as OD₅₃₀/OD₆₀₀/ml of culture supernatant (10). To test the effect of a denaturant on hemolytic activity, filtered culture supernatants were incubated with 6 M guanidine hydrochloride (GuHCl) for 15 min at room temperature. Enzyme assays on the GuHCl-treated supernatant was carried out as described above. The final concentration of GuHCl in enzymatic assays was 0.096 M.

Time course analysis of alpha-hemolysin secretion. Cultures were grown in LB containing ampicillin and chloramphenicol without IPTG to an OD₆₀₀ of ~0.3, at which point TolC synthesis was induced with 0.4 mM IPTG. A 5-ml aliquot was removed every 10 min for up to 60 min. Cells were pelleted by centrifugation in a bench-top centrifuge for 15 min at room temperature. The filtered supernatant was used to examine extracellular HlyA levels; TolC levels were determined from envelopes prepared from the pellet. HlyA and TolC levels were analyzed by Western blots.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Isolation of tolC mutants. Two genetic strategies were employed to isolate tolC mutants defective in alpha-hemolysin secretion. One strategy involved direct mutant selection in which a plasmid carrying the tolC gene was mutagenized to avoid mutations mapping in other genes that could also confer the desired phenotype. A mutator strain (mutD mutS mutT) was used for plasmid mutagenesis and a randomly mutagenized pool of tolC plasmids was transformed into a tolC::Tn10 strain containing an hly plasmid. Transformants were replica plated onto blood agar medium containing novobiocin and two additional antibiotics to select for the tolC and hly plasmids. The presence of novobiocin (10 µg/ml) selected against tolC null mutants, which are hypersensitive to this and other hydrophobic antibiotics. Mutants defective in alpha-hemolysin secretion produced smaller hemolytic zones compared to the tolC⁺ parental strain. Of the 18,000 transformants examined, 16 showed smaller hemolytic zones.

Protein and nucleotide sequence analyses. TolC was analyzed to see whether reduced protein levels resulted in the observed phenotype. Envelope proteins were transferred onto a membrane and hybridized with polyclonal antibodies raised against TolC and LamB (as an internal control). The data showed that only 10 of the 24 mutants produced a detectable level of TolC (Table 2). These 10 mutants were analyzed further. Nucleotide sequence of the entire tolC gene from 10 mutants was determined through employing primers complementary to the tolC sequence. This analysis revealed that 9 of the 10 isolates contained a single-base-pair substitution within a region of tolC encoding for the mature portion of the protein (Table 2). One mutant with an alteration affecting the last residue of the signal sequence was not analyzed further. Two mutants had an identical substitution (A343T), while two other isolates had different substitutions at the same site (A360T and A360V).

Phenotypic characterization of tolC mutants. None of the mutants showed hypersensitivity toward novobiocin (Table 2), which was expected since the isolation strategy demanded a proficient efflux activity. The slight increase in sensitivity observed in a mutant carrying a T434M substitution may have been due to the presence of reduced TolC levels in the envelope. The two mutants isolated through the colicin E1 selection method showed different degrees of sensitivity to colicin E1 in titration assays, with the S257P substitution resulting in complete insensitivity (Table 2). The other six mutants with substitutions in the mature portion of the TolC protein remained sensitive to colicin E1. tolC null mutants are known to be resistant to the bacteriophage TLS (9). All but one mutant, bearing a S257P substitution, were sensitive to TLS. Lastly, unlike the strain carrying a null tolC mutation, which drastically reduces OmpF levels (22, 24), all mutants isolated in this study showed normal OmpF expression (data not shown). These analyses showed that five mutants phenotypically resembled a strain expressing the parental TolC protein. The remaining three mutants bearing a L42P, S257P, or T434M substitution showed some degree of defect in at least one of the phenotypes associated with the tolC null mutation.

Secretion of alpha-hemolysin in tolC mutants. Mutants and control strains were tested for their abilities to secrete alpha-hemolysin by quantifying the amount of toxin secreted in the supernatant of cultures grown to late log phase. Bacterial cells were removed by centrifugation, supernatants were filtered through 0.45-µm (pore-size) cellulose acetate filters, diluted, and spotted onto a polyvinylidene difluoride membrane for dot blot analysis. Filters were hybridized with polyclonal antibodies raised against alpha-hemolysin, and hybridized spots were quantified (Table 3). The data showed that, with the exception of a mutant carrying a L42P substitution, all other mutants secreted alpha-hemolysin in amounts similar to that secreted by a strain expressing the parental TolC protein. Identical results were obtained when alpha-hemolysin precipitated by trichloroacetic acid from the culture supernatant was analyzed by Western blots (data not shown).

Hemolytic activities of tolC mutants. The cytolytic activity of secreted alpha-hemolysin from mutants and control strains was quantified in an attempt to resolve the above anomaly. The kinetics of blood lysis by culture supernatants were determined (Fig. 1), and relative values are presented in Table 3. The data revealed that the supernatants of all of the mutant cultures displayed significantly lower hemolytic activities than that observed in the parental strain. Based on their hemolytic activities, mutants can be divided into two groups. One group bearing a L42P, T140A, G147D, or S257P substitution showed dramatically lower activities ranging from 4 to 28% of the parental level. The second group, consisting of mutants with an A343T, A360T, A360V, or T434M substitution, had moderate activities ranging from 46 to 51% of the parental value. These results provided quantitative verification that the smaller hemolytic zones formed around colonies of mutant tolC strains were not due to reduced secreted alpha-hemolysin levels per se but rather to reduced hemolytic activities.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Cytolytic activities of secreted -hemolysin from TolC+, TolC, and TolC mutant strains. Graphs were obtained from three independent experiments in which the slope values did not vary more than 13% as shown in Table 3.

Restoration of hemolytic activity by GuHCl. One possibility for the presence of inactive alpha-hemolysin in the supernatant involves its aggregation or misfolding during secretion or postsecretion. A precedent for this notion exists in LPS mutants, in which the secretion of alpha-hemolysin into the medium was normal but secreted alpha-hemolysin molecules were enzymatically inactive due to aggregation (35). If alpha-hemolysin in the tolC mutants is aggregated or misfolded, it may be possible to resolve aggregates or restore correct folding by treatments with chaotropic agents such as GuHCl and restore hemolytic activity. Consistent with this notion, we discovered that the addition of 6 M GuHCl and its subsequent dilution to 0.1 M stimulated the hemolytic activity of toxin secreted from all but one strain expressing the TolC_G147D protein (Fig. 2). These results showed that the low hemolytic activities observed were likely due to the presence of misfolded or aggregated alpha-hemolysin in the medium.

View larger version (20K): [in this window] [in a new window]   FIG. 2.   Effect of GuHCl treatment on secreted alpha-hemolysin activity. Secreted alpha-hemolysin was treated with 6 M GuHCl () or untreated () as described in Materials and Methods.

Conformational alterations of alpha-hemolysin secreted from the mutant strain. We employed protease sensitivity assays, together with sedimentation and filtration analyses, to examine whether the inactivity of secreted alpha-hemolysin from the mutant was due to its aggregation or misfolding. Despite our repeated efforts with experiments involving limiting proteolysis with carboxypeptidase Y or trypsin, we failed to find any reproducible differences in the proteolytic pattern of alpha-hemolysin secreted from the mutant and parental strains.

View larger version (19K): [in this window] [in a new window]   FIG. 3.   Sedimentation analysis of alpha-hemolysin secreted from strains expressing either wild-type TolC () or TolCS257P (). The amounts of alpha-hemolysin remaining in the supernatant were determined and plotted after normalization to values obtained prior to centrifugation. See Results for details.

Time course analysis of alpha-hemolysin secretion. It is conceivable that an unproductive interaction between alpha-hemolysin and the mutant TolC protein impedes secretion, thus leading to misfolding and/or aggregation of the toxin molecules. To examine secretion, we exploited the dependency of alpha-hemolysin secretion on TolC whose synthesis from a plasmid was controlled by IPTG. In the absence of IPTG, only basal levels of TolC and alpha-hemolysin were detectable (Fig. 4). After the addition of IPTG, TolC levels rose rapidly; alpha-hemolysin was detected in the supernatant of a TolC⁺ culture at 10 min postinduction, and the levels rose steadily until 50 min postinduction.

View larger version (19K): [in this window] [in a new window]   FIG. 4.   Time course analysis of alpha-hemolysin secretion and TolC synthesis in strains expressing either wild-type TolC or TolCS257P proteins. The levels of alpha-hemolysin and TolC were determined after analyzing protein samples by use of Western blots. Further details of the experiment are provided in Materials and Methods.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

A goal of this research has been to obtain missense tolC mutations that primarily affect alpha-hemolysin secretion without displaying the pleiotropic phenotype of a more frequent class of null mutations. This was achieved by selecting tolC mutants on medium containing inhibitors that must be removed by cells expressing TolC proficient in its efflux activity. The desired class of tolC mutants defective in alpha-hemolysin secretion was identified among colonies that formed relatively small hemolytic zones on blood agar medium.

A unique property of TolC mutants. A novel class of tolC mutants emerged in which the steady-state level of secreted alpha-hemolysin was normal, but the secreted toxin was enzymatically less active. While the activity of extracellular alpha-hemolysin was reduced in the mutants, the cytoplasmic activity was identical to the parental strain, suggesting the toxin's inactivation during or after its secretion into the medium. Although all tolC mutants studied here showed reduced extracellular hemolytic activity, this defect was particularly pronounced in two mutants expressing TolC_T140A or TolC_S257P. In these mutants, TolC and secreted alpha-hemolysin levels were similar to that of the parental strain, but the extracellular hemolytic activity was 10-fold reduced.

Restoration of hemolytic activity. In several instances, the treatment of inactive extracellular alpha-hemolysin with GuHCl substantially elevated its cytolytic activity. This increase in hemolytic activity likely reflected a decrease in aggregation and/or misfolding. Interestingly, a modest improvement of hemolytic activity was also observed in a strain expressing the wild-type TolC protein, indicating that a small population of secreted alpha-hemolysin molecules may exist as inactive aggregates. Indeed, the inactivation of secreted alpha-hemolysin due to aggregation in the aqueous environment is thought to be a normal phenomenon (11, 29). Since GuHCl restored the hemolytic activity in vitro, it appeared unlikely that the step of HlyC-mediated acylation, which is essential for the cytolytic activity (36), was affected in tolC mutants. Consistent with this notion, the cytoplasmic hemolytic activity in the mutant strains was found to be identical to that of the parental strain (data not shown).

tolC and rfa mutants display similar alpha-hemolysin activity defects. Two other studies have shown the presence of predominantly inactive alpha-hemolysin molecules in the culture supernatant of mutant E. coli strains. In these cases, however, the genetic defect existed within the rfaC (1) and rfaP (35) genes whose products influence the composition of LPS inner core (32). rfaC and rfaP mutants display a pleiotropic phenotype, including hypersensitivity toward hydrophobic compounds and defective outer membrane protein biogenesis (28). Stanley et al. (35) demonstrated that aggregation of alpha-hemolysin in the culture supernatant resulted in the inactivation of its hemolytic activity. In the rfaC mutant, both the extracellular expression and activity of alpha-hemolysin was affected (1). These studies suggested that the intact LPS core interacts with alpha-hemolysin either during or after its secretion, and this interaction affords a conformational stability to the toxin molecule. Unlike the rfaC and rfaP mutants, the tolC mutants studied here do not display a severe membrane defect, as was evident from the lack of a hypersensitivity phenotype. Thus, it appears unlikely that the effect of mutant TolC proteins on the activity of alpha-hemolysin is indirect and produced through a gross LPS or membrane defect. Polyacrylamide gel analysis of LPS from envelopes of TolC mutants showed no qualitative or quantitative difference from LPS of the parental strain (data not shown), thus further supporting the notion that the reduced hemolytic activity in TolC mutants is not due to an effect on LPS.

Possible reasons for alpha-hemolysin inactivation. Due to the relatively large size of alpha-hemolysin, it is likely to translocate through the TolC barrel in a partially unfolded state. Subsequent to its release into the medium or through its interaction with the target cell, the alpha-hemolysin molecule must then refold in order to become enzymatically active. In the TolC mutants studied here, this step of refolding might be defective. However, since not all substitutions map within the -barrel domain of TolC, it is difficult to envisage how the terminal steps of alpha-hemolysin translocation, i.e., release and refolding, would be affected in the mutants. Time course studies involving alpha-hemolysin secretion in a mutant strain expressing TolC_S257P showed (i) a delayed appearance of toxin in the supernatant and (ii) a slower rate of secretion compared to that observed in the parental strain. These observations suggest that the steps involving alpha-hemolysin translocation through the TolC barrel might be affected in the mutant strains. A block in translocation may lead to misfolding and/or aggregation of toxin molecules trapped within the TolC barrel.

View larger version (50K): [in this window] [in a new window]   FIG. 5.   Three-dimensional structure of TolC showing locations of various mutant substitutions that influence alpha-hemolysin secretion and activity. Views from the side and proximal end are shown. Colored circles reflect the location of various mutant sites (in space-filled representations) in -helices (H), -strands (S), or turns (T).