- Email: [email protected]
Need for TolC, anEscherichia coliouter membrane protein, in the secretion of heat-stable enterotoxin I across the outer membrane
Microbial Pathogenesis 1998; 25: 111–120 Article No. mi980211
MICROBIAL PATHOGENESIS
Need for TolC, an Escherichia coli outer membrane protein, in the secretion of heat-stable enterotoxin I across the outer membrane Hiroyasu Yamanakaa, Tomohiko Nomuraa, Yoshio Fujiib & Keinosuke Okamoto∗c a
Department of Biochemistry, Faculty of Pharmaceutical Sciences, and bInstitute of Pharmacognosy, Tokushima Bunri University, Yamashiro, Tokushima, Tokushima 770-8514, Japan (Received April 13, 1998; accepted in revised form April 30, 1998)
Escherichia coli heat-stable enterotoxin Ip (STIp) is a typical extracellular toxin consisting of 18 amino acid residues synthesized as a precursor of pre (amino acid residues 1 to 19), pro (amino acid residues 20 to 54), and mature (amino acid residues 55 to 72) regions. STIp synthesized in the cytoplasm must cross the inner and outer membranes to migrate into the extracellular environment. Previous studies showed that the precursor translocates across the inner membrane utilizing the general export pathway consisting of Sec proteins. However, it remains unclear how it crosses the outer membrane. In this study, we examined the effects of mutation of the tolC gene which encodes an E. coli outer membrane protein, TolC, on the release of STIp into the extracellular environment. The mutation reduced the amount of STIp released into culture supernatant and increased the amount of STIp accumulated in the periplasm. This indicates that TolC mediates the translocation of STIp across the outer membrane. The inability to transfer STIp in the periplasm into the culture supernatant was restored by introduction of the tolC gene into the mutant cells. In the mouse intestinal loop assay, living cells of the mutants did not show a positive response, but wild-type cells did. These results showed that TolC is involved in the translocation of STIp 1998 Academic Press across the outer membrane. Key words: Escherichia coli, outer membrane, enterotoxin, TolC, secretion.
Introduction Exoproteins produced by gram-negative bacteria have to cross the cell envelope, which is ∗ Author for correspondence. 0882–4010/98/090111+10 $30.00/0
composed of two membranes (inner and outer membranes) separated by an aqueous periplasmic compartment containing the peptidoglycan cell wall [1, 2]. To date, three secretory pathways have been recognized in gram-negative bacteria; type I, type II and type III pathways [3–5]. Exoproteins employing the type I 1998 Academic Press
112
pathway are secreted across the bacterial envelope in a single step mediated by ABC (ATPbinding-cassette) transporter [6]. On the other hand, extracellular proteins secreted by the type II pathway are synthesized as precursor proteins with an amino-terminal signal sequence and are translocated across the inner membrane via the Sec protein machinery. The proteins released into the periplasm then cross the outer membrane mediated by an auxiliary secretion machinery composed of 13 to 15 accessory proteins [4, 7, 8]. The type III secretion system has been more recently identified and it has yet to be characterized in detail [3]. Escherichia coli heat-stable enterotoxin Ip (STIp) is an extracellular peptide toxin produced by enterotoxigenic E. coli strains [9]. The toxin is synthesized as a precursor consisting of three regions; pre (amino acid residues 1 to 19), pro (amino acid residues 20 to 54) and mature (amino acid residues 55 to 72) regions [10, 11]. Previous studies demonstrated that translocation of the STIp precursor across the inner membrane is achieved by the general export machinery consisting of Sec proteins, and the pre region of STIp functions as a signal peptide in translocation. After or during translocation across the inner membrane, the pre region is cleaved and the peptide generated, consisting of the pro and mature regions, is released into the periplasm [10]. Subsequently, we found that cleavage between the pro region and the mature region, and intramolecular disulfide bond formation in the mature region occur in the periplasmic space and that the mature STIp molecules thus generated cross the outer membrane and emerge in the culture supernatant [12, 13]. However, it remains unclear how the mature STIp molecules cross the outer membrane. As described above, the type II exoproteins released into the periplasmic space utilize the auxiliary secretion machinery composed of 13 to 15 accessory proteins to cross the outer membrane [4]. As the genes for these accessory proteins are not included in the normal E. coli chromosome, the type II exoproteins are not released into the culture supernatant of E. coli strains when only the protein gene (which does not contain the genes encoding the accessory proteins) are introduced into the cells [4]. However, STI is released into the culture supernatant from the E. coli transformed with plasmids encoding the STI gene but not any accessory protein genes [11]. Thus, the secretion machinery necessary for the translocation of STI across the
H. Yamanaka et al.
outer membrane is present in normal E. coli strains. The secretion therefore appears to be an intrinsic pathway in E. coli. It has been shown that TolC, an E. coli outer membrane protein [14], is involved in translocation of haemolysin [15] and colicin V [16] across the outer membrane. Furthermore, it has been reported that TolC is also related to the efflux of small molecules such as antibiotics [17, 18]. Therefore, we suspected that TolC is also involved in the translocation of STI across the outer membrane.
Results Disruption of the tolC gene of E. coli BL21 Strain BL21 has advantage in expression of recombinant proteins in E. coli, as the strain had lost both lon and ompT protease [19, 20]. Moreover, BL21 can be used as the host strain in the T7 expression system, as it contains T7 RNA polymerase gene on the chromosome [20]. We constructed a tolC-deficient mutant of BL21 by transduction with the P1 phage [21]. Escherischia coli CS1562 (k−, tolC6::Tn10) [22] was used as the parent strain for the mutation. After transduction with P1 phage, the cells grown in the presence of tetracycline (50 lg/ml) (Tn10 encodes the tetracycline resistance gene) were examined for sensitivity to SDS and colicin E1, characteristic properties accompanying tolC mutation as reported previously [23]. The cells that were sensitive to SDS and resistant to colicin E1 were selected as candidate tolC-deficient mutants. The tolC gene of the candidates was examined by PCR. The primers used were (i) and (ii). The sequence of the primer (i) is 5′ TTAAATGTCCTGGCACTAATA 3′ (sense) and that of (ii) is 5′ TGCGGCAGATAACCCGTATC 3′ (antisense). These extend from bases 101 to 120 and 1831 to 1850 of the tolC gene sequence reported by Niki et al. [24], respectively. A DNA fragment of the expected length (1750 bp) was amplified from the DNAs of wild-type strain (BL21), but no fragment was amplified from the DNAs of candidate cells (data not shown), suggesting that the tolC was mutated in these cells. To confirm the mutation of tolC, the production of the tolC mRNA from these strains was examined. Total cellular RNA was purified from the cells grown in L broth [25] at 37°C to
TolC and translocation of E. coli STI
exponential phase by the method described by MacDonald et al. [26]. The tolC mRNA in the sample was enzymatically amplified by reverse transcription reaction and PCR. The RT-PCR kit (R019A, Takara Shuzo Co., Kyoto, Japan) was used for the reaction. The sequences of the primers used in this experiment were (iii) 5′ TAACGCTGCAGGAAAAAGCA 3′ (sense) and (iv) 5′ TGCGGCAGATAACCCGTATC 3′ (antisense), extending from bases 618 to 637 and 1831 to 1850 of the tolC gene sequence reported by Niki et al. [24], respectively. The predicted length of DNA fragment amplified by the reaction is 1233 base pairs. The amplified DNAs were analysed by acrylamide gel electophoresis. The DNA fragment of 1233 base pairs was amplified from the sample of BL21, as expected, however, the fragment was not detected in the sample from the candidate cells (data not shown). This shows that tolC was mutated and was not transcribed in the candidate strain. The tolC mutant strain was named BL21-2.
Effect of tolC mutation on release of STI from cells To examine the effects of tolC mutation on the release of STI into culture supernatant, BL21 (tolC+) and BL21-2 (tolC−) were transformed with Tc-1 carrying STIp gene (Table 1). The transformants were cultured in L broth containing ampicillin (50 lg/ml) at 37°C with shaking for 5 h. The cells were separated from the culture supernant by centrifugation. The periplasmic fraction of the cells was prepared as described in Materials and methods and the volume of the fraction was adjusted to make it equal with that of the culture supernatant. The STI activity of these samples was measured using the suckling mouse assay. As shown in Fig. 1(a), the STI activity of the culture supernatant of BL21/Tc-1 was high, but that of BL21-2/Tc-1 was extremely low. In contrast, STI activity of the periplasmic fraction of BL21/Tc-1 was low and that of BL21-2/Tc-1 was high [Fig. 1(b)]. This indicated that STI produced in the tolC mutant cells did not translocate across the outer membrane of the mutant cells and accumulated in the periplasmic space.
Complementation of the tolC mutation To confirm the involvement of TolC in the translocation of STI, we examined whether the release
113
of STI into culture supernatant in the tolC mutant strain was restored by introducing the wild-type tolC gene into the tolC mutant. To carry out this experiment, Tc-1-TolC plasmid was constructed by inserting the tolC gene into Tc-1. As a control, the mutant tolC gene which does not produce functional TolC was inserted into Tc-1 as described in Materials and methods. The obtained plasmid carrying the mutant tolC gene was named Tc-1-mTolC (Table 1). The tolC mutant (BL21-2) was transformed with these plasmids and the transformed cells were cultured in L broth as described above. STI activity of the culture supernatant and of the periplasmic fraction of each strain was determined. The level of STI activity of culture supernatant of BL21-2/Tc-1-TolC was higher than that of the strain transformed with Tc-1 (BL21-2/Tc-1) and was close to that in wildtype cells (BL21/Tc-1) [Fig. 1(a)]. In contrast, the remaining STI activity in the cells of BL21-2/ Tc-1-TolC was very low [Fig. 1(b)]. These results indicated that the reduced efficiency of STI to cross the outer membrane caused by mutation of the tolC gene was restored by introducing the wild-type tolC gene into the mutant strain. However, the reduced efficiency was not restored by introducing the plasmid carrying the mutant tolC gene (Tc-1-mTolC) (Fig. 1). This further supports the notion that TolC is involved in the translocation of STI across the outer membrane.
Analysis by pulse-labelling The above results indicated that TolC is needed for STI to cross the outer membrane. However, the results were obtained by determination of STI activities of E. coli cells after long-term cultivation. It was possible that the STI activity of the mutant was affected by the side effects of the tolC mutation during long-term cultivation. Such side effects may lead to wrong conclusions. Analysis of the nascent STI may provide more accurate information regarding the effects of mutation of the tolC gene. We therefore analysed the nascent STI by pulse-labelling using pET system. pET11-STI plasmid contains STIp gene. pET11-STI-TolC plasmid and pET11-STI-mTolC plasmid were constructed by inserting the wild tolC gene and the mutant tolC gene (which does not produce functional TolC), respectively, into pET11-STI
114
H. Yamanaka et al.
Table 1. Strains and plasmids Strain or plasmid Strains E. coli BL21 CS1562 2086 BL21-2 Plasmids pBR322 Tc-1 pET11 pT7-Blue(R) pT7-TolC pT7-mTolC Tc-1-TolC Tc-1-mTolC pET11-STI pET11-STI-TolC pET11-STI-mTolC
Features
Source or reference
Containing T7 RNA polymerase gene on the chromosome E. coli K-12 derivative strain (F−, k−, tolC::Tn10) Clinical isolate producing STI TolC mutant of BL21
[20]
Vector plasmid Derivative of pBR322 containing STIp gene (parental plasmid for STIp gene) Containing a T7 promoter under the control of the lac operator Cloning vector for PCR-amplified gene Derivative of pT7-Blue(R) containing tolC gene Introduction of frame shift mutation at PstI site in the tolC gene of pT7-TolC Insertion of the tolC gene into BamHI site of Tc-1 Insertion of the mutant tolC gene encoded in pT7mTolC into BamHI site of Tc-1 Insertion of STIp structural gene of Tc-1 into NdeI site of pET11 under the control of the T7 promoter Insertion of tolC gene into HindIII-EcoRI site of pET11-STI Insertion of the mutant tolC gene encoded in pT7mTolC into HindIII-EcoRI site of pET11-Tc-1
(Table 1). BL21 and BL21-2 which were transformed with these pET11 derivatives were pulselabelled with l-[35S]cysteine for 3 min and chased for 3 min. BL21 cells harbouring pET11, which does not encode the STI gene, were used as a negative control. The volume of these cultures was 1 ml. The culture supernatant and the periplasmic fractions of the cultures were prepared as described in Materials and methods. The volume of each periplasmic fraction was adjusted to 0.4 ml. The samples prepared were resolved by SDS-PAGE. The samples from the culture supernatant and the periplasmic fraction were run in the gels and shown in Fig. 2(a) and (b), respectively. In the gel shown in Fig. 2(a), a dense band appeared in lanes 2 and 4 at a position corresponding to that of the mature STIp comprised of 18 amino acid residues [13]. This indicated that the culture supernatant from the wild-type strain (BL21/pET11-STI) contained STIp [lane 2, Fig. 2(a)], but that from the tolC mutant strain (BL21-2/pET11-STI) did not [lane 3, Fig. 2(a)].
[22] Our collection This study Takara Shuzo Co. [11] [20] Takara Shuzo Co. This study This study This study This study [13] This study This study
However, the culture supernatant from the mutant strain which was complemented with the native tolC gene (BL21-2/pET11-STI-TolC) contained STIp [lane 4, Fig. 2(a)]. Recovery of STI production was not achieved by introduction of the mutant tolC gene [lane 5, Fig. 2(a)]. Analysis of periplasmic samples showed that the density of the band corresponding to mature STIp (indicated by the arrow) was faint in the wild-type strain [BL21/pET11-STI, lane 2 of Fig. 2(b)]. This indicated that most mature STIp had already moved into the culture supernatant in the wild-type strain. However, the band in Fig. 2(b), lane 3, was dense. This sample was prepared from the periplasmic fraction of tolC mutant strain (BL21-2/pET11-STI). As shown in lane 3 of Fig. 2(a), the amount of STI released into culture supernatant from the mutant cells was very low. Thus, the STI which emerged into the periplasmic space of the mutant cells could not pass through the outer membrane. This accumulation of STI in the periplasmic space did not occur in the mutant strain complemented
TolC and translocation of E. coli STI
25
BL21/Tc-1
BL21-2/Tc-1
BL21-2/Tc-1-TolC
BL21/Tc-1-mTolC
BL21/Tc-1
BL21-2/Tc-1
BL21-2/Tc-1-TolC
BL21/Tc-1-mTolC
50
BL21/pBR322
Enterotoxic activity of tolC mutant
0
STI activity (unit)
the mutant tolC gene [lanes 5 of Figs 2(a) and (b)]. These results support the notion that TolC is directly involved in the translocation of STI across the outer membrane.
(a)
BL21/pBR322
STI activity (unit)
50
115
(b)
25
0
Figure 1. Effects of mutation of the tolC gene on STI activity of cells. BL21 (tolC-positive strain) and BL212 (tolC-negative strain) were transformed with plasmids Tc-1, Tc-1-TolC and Tc-1-mTolC carrying STIp, STIp and tolC, and STIp and the mutant tolC genes, respectively, as shown in Table 1. The transformed cells were cultured in L broth containing ampicillim (50 lg/ml) at 37°C for 5 h with shaking. After cultivation, the culture supernatants (a) and the periplasmic fractions (b) were prepared as described in the text. Enterotoxic activities of these samples were determined by the suckling mouse assay.
with the native tolC gene [BL21/pET11-STI-TolC, lane 4 of Fig. 2(b)] and most STI translocated across the outer membrane [lane 4, Fig. 2(a)]. However, the ability of the mutant cells to transport the STI in the periplasmic space into culture supernatant was not restored by introduction of
To clarify further the role of TolC in the translocation of STI across the outer membrane, the enterotoxic activity of tolC mutant cells carrying STIp gene was examined. BL21-2/Tc-1 and BL21/Tc-1 were each cultured on nutrient agar at 37°C for 6 h. The cells cultured were collected, washed with 0.5% NaCl and then suspended in L broth at a concentration of 1.5×107 colonyforming units (cfu)/ml. A portion of the cell suspension (0.2 ml) was injected into a mouse intestinal loop. Three hours later, the amount of fluid accumulated in the loop was determined. As shown in Fig. 3, BL21/Tc-1 induced a marked positive response {fluid accumulation ratio [weight of loop (g)/length of loop (cm)] above 0.20 was regarded as a positive response}, while BL21-2/Tc-1 did not. This indicated that sufficient amounts of STI for accumulation of the fluid in the intestinal lumen were not released from the tolC-negative strain (BL21-2/Tc-1), although the amount of STI released from the tolC-positive strain (BL21/Tc-1) was sufficient. These observations indicated that TolC is important for release of STI not only in cells cultured in vitro but also in those inoculated into the intestinal lumen of mice.
Discussion We analysed the translocation of STI across the outer membrane of E. coli and found that TolC is an indispensable element in transport of STI from the periplasmic space to the culture supernatant in BL21. However, it remained unclear whether TolC functions in clinical isolates as well as in the strains using for gene manipulation. As described above, BL21 had lost both lon and ompT proteases. The decomposition rate of proteins synthesized is therefore thought to be slower in BL21 compared with clinical isolates. Furthermore, the plasmid encoding the STI gene used in this experiment was a multicopy plasmid. Thus, the amounts of STI synthesized in strains harbouring the plasmid would be
1
2
3
4
BL21-2/pET11-STI-mToIC
BL21-2/pET11-STI-TolC
(b)
5
5
BL21-2/pET11-STI-mToIC
BL21-2/pET11-STI
4
BL21-2/pET11-STI-TolC
3
BL21-2/pET11-STI
2
BL21/pET11-STI
1
BL21/pET11
(a)
BL21/pET11-STI
H. Yamanaka et al.
BL21/pET11
116
higher than that in clinical isolates. Therefore, the STI secretion pathway in the strain used in this study may have been different from that of clinical isolates, and therefore the conclusions guided from the results of this experiment could only be applied to the BL21 strain harbouring multicopy plasmid. To confirm the role of TolC in clinical isolates, we mutated the tolC gene of several clinical isolates of E. coli producing STI by P1 transduction and examined the amounts of STI released into the culture supernatant of the mutant cells. The amount was low in all mutants (data not shown), showing that TolC is important for the release of STI into the culture supernatant not only in BL21 but also in the clinical isolates. Therefore, it can be concluded that TolC is needed for STI to cross the outer membrane in all E. coli strains. TolC has been reported to be involved in the secretion of both haemolysin and colicin V [15, 16]. These proteins pass through the inner and outer membranes in an all-or-nothing fashion; that is, the proteins synthesized in the cytoplasm utilize a secretory machinery consisting of a translocator in the inner membrane and of TolC in the outer membrane. The inner membrane translocator is comprised of an ABC exporter and an accessory protein [27]. It is said that the accessory protein acts as a bridge to interact with TolC in outer membrane [18]. As the components comprising the machinery are connected, proteins which utilize the secretory machinery translocate across the two
Figure 2. Effects of mutation of the tolC gene on translocation of STI synthesized in cells across the outer membrane. To selectively express STI, the T7 expression system was used. BL21 (tolC-positive strain) and BL21-2 (tolC-negative strain) were used as host strains. These host strains were transformed with derivatives of pET11 (Table 1). The transformed cells were cultured in L broth until reaching the exponential phase. After induction with 2 mM IPTG, the cells were labelled for 3 min with l-[35S]cysteine (10 lCi). The labelling reaction was terminated by adding cold cysteine into the cultures, and incubation was continued for 3 min (chase period). After the chase period, the culture supernatants and the periplasmic fraction of cells were prepared as described in the text. The samples were resolved by SDSPAGE and radioactive bands were detected by image analysis. The arrow indicates the position of the mature STIp. (a) culture supernatant; (b) periplasmic fraction.
TolC and translocation of E. coli STI
117
Postive response
0.2
Negative response
Fluid accumulation (g/cm)
0.3
BL21-2/Tc-1
BL21/Tc-1
0.1
Cells inoculated
Figure 3. Enterotoxic activities of tolC mutants. BL21 (tolC-positive strain) and BL21-2 (tolC-negative strain) were transformed with Tc-1 carrying the STIp gene. The cells which were grown on nutrient agar were suspended in L broth at a concentration of 1.5×107 cfu/ml. A portion of the cell suspension was injected into a mouse intestinal loop. Three hours later, the amount of fluid accumulated in the loop was determined. Fluid accumulation ratio [weight of loop (g)/length of loop (cm)] above 0.20 was regarded as a positive response.
membranes without stopping. Therefore, the protein either crosses both membranes or remains in the cytoplasm, with no intermediate protein detected in the periplasm. In contrast, STI translocates across the inner membrane in a SecA-dependent manner, and intermediates of STI have been detected in the periplasm [12, 13]. The results of the present study showed that TolC is involved in the movement of STI from the periplasm to the extracellular environment. As TolC in the outer membrane forms a trimer [14], STI might pass through the pore constructed by association of three molecules of TolC.
As described above, TolC is regarded as a component of the secretory machinery and is thought to be involved in the translocation of protein through both inner and outer membranes in an all-or-nothing fashion. The present study showed that TolC also functions in the translocation of proteins brought into the periplasmic space by Sec proteins across the outer membrane. From these results, it can be concluded that TolC functions in translocation across the outer membrane not only of proteins which pass through the two membranes in an all-or-nothing fashion but also of proteins which are carried into the periplasmic space in a SecAdependent manner. Such a notion about TolC was first proposed by Foreman et al. [28]. They examined the secretion of STII, another heat-stable enterotoxin of E. coli, in which the amino acid sequence is quite different from that of STI, into culture supernatant and found that the efficiency of the release of STII into culture supernatant was low in a tolC mutant strain. As the translocation of STII across the inner membrane is achieved by the machinery consisting of TolC from the result. Recently, we also examined the effect of the mutation of the tolC gene on the secretion of STII into culture supernatant by the system used in this study. The result also showed that the efficiency of translocation of STII across the outer membrane was reduced by the tolC mutation. However, a small amount of STII was released from the mutant cells (unpublished data). This indicates that the mutation of tolC affects the secretion of STII. However, it is not elucidated that TolC is directly involved in the translocation of STII across the outer membrane. Further studies are necessary to clarify the role of TolC in the secretion of STII. The mechanism by which TolC recognizes STI in the periplasm might be quite different from those involved in the recognition of the haemolysin and colicin V. Recent studies showed that TolC is also involved in the efflux of small toxic components including antibiotics [17]. Therefore, studies of the recognition of STI by TolC might be useful to gain insight into the characteristics of the antibiotic-resistant bacteria.
Materials and methods Bacterial strains and plasmids The E. coli strains used in this study are listed in Table 1. E. coli CS1562 (k−, tolC6::Tn10) [22],
118
kindly provided by B. J. Backmann (E. coli Genetic Stock Center, Yale University, New Heaven, Conn., U.S.A.), was used as the tolC mutant parental strain. Plasmid Tc-1 encoding the STIp gene was the wild-type plasmid [11]. To selectively express STIp, the pET system was used [20]; in pET11STI, the STIp gene was cloned into NdeI of pET11 bringing its production under control of the T7 promoter [13].
Cloning and mutation of tolC gene The DNA fragment of the tolC gene was cloned as follows. The tolC gene was amplified from the chromosomal DNA of an STI-producing E. coli clinical isolate by PCR (strain 2086, Table 1). The primers used are (i) and (ii) which are described in Results. The DNA fragments amplified with these primers encompassed both the promoter and the structural regions of the tolC gene. The fragments were ligated into pT7Blue(R) (Takara Shuzo Co.), a cloning vector for PCR-amplified DNA fragments. The plasmid obtained was named pT7-tolC (Table 1). The mutant tolC gene was prepared as follows, pT7-TolC has two PstI sites, one in the tolC gene and the other in the multi-cloning site of pT7Blue(R). The nucleotide sequence at the position which is recognized by PstI in the tolC gene encodes the amino acid residue at position 107 from the amino terminus of TolC [24]. As TolC is composed of 495 amino acid residues, this frame shift mutation at the PstI site results in production of a mutant protein which is nonfunctional. To introduce the frame shift mutation, pT7-TolC was partially digested with PstI, and the resultant linear fragment was purified. The fragment was treated with Klenow to digest the protruding terminus at 3′ single-stranded regions, and the fragments generated were ligated. Then, the plasmid containing a frame shift mutation at the PstI site in the tolC gene was selected. The plasmid obtained was designated pT7-mTolC (Table 1).
Insertion of tolC into plasmids carrying the STI gene Insertion of the tolC gene into Tc-1 was carried out as follows: pT7-TolC was digested with HindIII and BamHI and the resulting linear fragment containing tolC gene was isolated. The
H. Yamanaka et al.
fragment was treated with Klenow to fill in the 5′ single-stranded ends. The fragment was cloned into the Klenow-treated BamHI site of Tc-1. The plasmid obtained was designated Tc1-TolC. Similarly, the mutant tolC gene isolated from pT7-mTolC was inserted into Tc-1, and the plasmid obtained was designated Tc-1-mTolC (Table 1). Insertion of the tolC gene into pET11-STI was carried out as follows: pT7-TolC was digested with HindIII and EcoRI and the resulting linear fragment containing tolC gene was isolated. The fragment was ligated to the HindIII-EcoRI fragment of pET11-STI encoding the STIp gene. The plasmid obtained was designated pET11-STITolC. Similarly, the mutant tolC gene isolated from pT7-mTolC was inserted into pET11-STI, and the plasmid obtained was designated pET11-STI-mTolC (Table 1).
Pulse-labelling BL21 cells transformed with pET11 derivative plasmids were shaken at 37°C until they reached an optical density at 660 nm of 0.4 in L broth containing ampicillin (50 lg/ml). The cultures were centrifuged, and the cell pellets were suspended in M9 minimal medium [25] containing an amino acid mixture (final concentration of each amino acid, 50 lg/ml) without cysteine at an optical density at 660 nm of 3.0. Afer adding a final concentration of 2 mM isopropyl-b-dthiogalactopyranoside (IPTG), the suspension was shaken at 37°C for 20 min. A portion (1 ml) was labelled for 3 min at 37°C with 10 lCi of l[35S]cysteine (0.5 to 1 kCi/mmol, ARS-101: American Radiolabeled Chemicals, Inc., St Louis, MO, U.S.A.) per ml. Incorporation of the label was terminated by adding 250 lg of unlabelled cysteine. The incubation was continued for 3 min. Thereafter, the supernatant was separated from the cells by centrifugation. The periplasmic fraction of the cells was prepared as described below. Prepared samples were mixed with SDS dye solution [29], heated at 100°C for 10 min and resolved by SDS-20% polyacrylamide gel electrophoresis (PAGE) [29]. The radioactive bands were visualized by autoradiography using an imaging plate analyzer (BAS 2000, Fujix, Tokyo, Japan).
Preparation of the periplasmic fraction Cells cultured in liquid medium were collected by centrifugation at 12 000 g for 10 min, then the
TolC and translocation of E. coli STI
cell pellet was suspended in 10 mM Tris-HCl buffer (pH 7.5). To prepare the periplasmic fraction, the cell suspension was incubated with polymyxin B at a concentration of 6500 U/mol at 4°C for 15 min and centrifuged (12 000 g for 15 min). The supernatant obtained was used as the periplasmic fraction.
Assay of STI activity STI activity was determined by means the suckling mouse assay as described [30]. Briefly, 0.1ml samples were administered via a gastric tube into the stomach of 2–3-day-old suckling mice, using 0.01% Evans blue due as a marker. The mice were killed 3.0 h later, and a ratio of intestinal to body weight of over 0.083 was considered positive. One unit was defined as the minimum effective dose that elicited a positive response, and the enterotoxin titre of samples was expressed as the reciprocal of the highest dilution that gave 1 unit of enterotoxin activity per 0.1 ml. The enterotoxic activity of each sample was determined in five mice.
Mouse intestinal loop assay Mice weighing 30–35 g were anaesthetized with sodium pentobarbital, and the intestines were exteriorized through a midline incision. The intestinal lumen was rinsed three times with saline. After rinsing, a series of ligated intestinal segments (loops), about 3 cm long and separated by a 0.5 to 1 cm interloop, were created. The most proximal loop was placed about 3 cm distal to the ligament of Treitz. One or two loops were created per intestine, and each loop was injected with 0.2 ml of sample solution. Three hours after injection, the mice were killed and the sample activities were measured and expressed as the ratio of the weight of the loop (in g) to its length (in cm). A ratio of over 0.20 was regarded as a positive response.
Acknowledgement This work was supported in part by a grant-in-aid for scientific research from the Ministry of Education, Science and Culture, Japan.
119
References 1 Nikaido H. Outer membrane. In: Neidhardt FC, Curtiss III R, Ingraham JL, Lin ECC, Low Jr. KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE, Eds. Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd edn. Waashington, D.C.: American Society for Microbiology, 1996; 29–47. 2 Wandersman C. Secretion across the bacterial outer membrane. In: Neidhardt FC, Curtiss III R, Ingraham JL, Lin ECC, Low Jr. KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE, Eds. Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd edn. Washington, D.C.: American Society for Microbiology, 1996; 955–966. 3 Lee CA. Type III secretion systems: machines to deliver bacterial proteins into eukaryotic cells? Trends Microbiol 1997; 5: 148–56. 4 Pugsley AP. The complete general secretory pathway in Gram-negative bacteria. Microbiol Rev 1993; 57: 50–108. 5 Salmond GPC, Reeves PJ. Membrane traffic wardens and protein secretion in gram-negative bacteria. TIBS 1993; 18: 7–12. 6 Fath MJ, Kolter P. ABC transporters: bacterial exporters. Microbiol Rev 1993; 57: 995–1017. 7 Arkowitz RA, Bassilana M. Protein translocation in Escherichia coli. Biochim Biophys Acta 1994; 1197: 311–43. 8 Den Blaauwen T, Driessen AJM. Sec-dependent preprotein translocation in bacteria. Arch Microbiol 1996; 165: 1–8. 9 Takao T, Hitouji T, Aimoto S, et al. Amino acid sequence of a heat-stable enterotoxin from enterotoxigenic Escherichia coli 18D. FEBS Lett 1983; 152: 1–5. 10 Okamoto K, Takahara M. Synthesis of Escherichia coli heat-stable enterotoxin STp as a pre-pro form and role of the pro sequence in secretion. J Bact 1990; 172: 5260–5. 11 So M, McCarthy BJ. Nucleotide sequence of bacterial transposon Tn 1681 encoding a heat-stable (ST) toxin and its identification in enterotoxigenic Escherichia coli strains. Proc Natl Acad Sci USA 1980; 77: 4011–15. 12 Yamanaka H, Kameyama M, Baba T, Fujii Y, Okamoto K. Maturation pathway of Escherichia coli heat-stable enterotoxin I: requirement of DsbA for disulfide bond formation. J Bact 1994; 176: 2906–13. 13 Yamanaka H, Nomura T, Fujii Y, Okamoto K. Extracellular secretion of Escherichia coli heat-stable enterotoxin I across the outer membrane. J Bact 1997; 179: 3383–90. 14 Koronakis V, Li J, Koronakis E, Stauffer K. Structure of TolC, the outer membrane component of the bacterial type I efflux system, derived from two-dimensional crystals. Mol Microbiol 1997; 23: 617–26. 15 Wandersman C, Delepelaire P. TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc Natl Acad Sci USA 1990; 87: 4776–80. 16 Gilson L, Mahanty HK, Kolter R. Genetic analysis of an MDR-like export system: the secretion of colicin V. EMBO J 1990; 9: 3875–84. 17 Fralick JA. Evidence that TolC is required for functioning of the Mar/AcrAB efflux pump of Escherichia coli. J Bact 1996; 178: 5803–5. 18 Paulsen IT, Park JH, Choi PS, Saier M. A family of gram-negative bacterial outer membrane factors that function in the export of proteins, carbohydrates, drugs
120
19 20 21 22
23
24
and heavy metals from gram-negative bacteria. FEMS Microbiol Lett 1997; 156: 1–8. Phillips TA, VanBogelen RA, Neidhardt FC. lon gene product of Escherichia coli is a heat-shock protein. J Bact 1984; 159: 283–7. Studier FW, Rosenberg AH, Dunn JJ, Dubendorff JW. Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol 1990; 185: 60–89. Silhavy TJ, Berman ML, Enquist LW. Experiments with Gene Fusions. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory, 1984. Austin EA, Graves JF, Hite LA, Parker CT, Schnaitman CA. Genetic analysis of lipopolysaccharide core biosynthesis by Escherichia coli K-12: insertion mutagenesis of the rfa locus. J Bact 1990; 172: 5312–25. Binet R, Wandersman C. Cloning of the Serratia marcescens hasF gene encoding the has ABC exporter outer membrane component: a TolC analogue. Mol Microbiol 1996; 22: 265–73. Niki H, Imamura R, Ogura T, Hiraga S. Nucleotide
H. Yamanaka et al.
25 26 27 28
29 30
sequence of the tolC gene of Escherichia coli. Nucl Acids Res 1990; 18: 5547. Miller JH. Experiments in Molecular Genetics. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory, 1972. MacDonald RJ, Swift GH, Przybyla AE, Chirgwin JM. Isolation of RNA using guanidinium salts. Methods Enzymol 1987; 152: 219–27. Higgins CF. ABC transporters: from microorganisms to man. Ann Rev Cell Biol 1992; 8: 67–113. Foreman D, Martinez I, Coombs G, Torres A, Kupersztoch YM. TolC and DsbA are needed for the secretion of STb, a heat-stable enterotoxin of Escherichia coli. Mol Microbiol 1995; 18: 237–45. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227: 680–5. Okamoto K, Okamoto K, Yukitake J, Miyama A. Reduction of enterotoxic activity of Escherichia coli heatstable enterotoxin by substitution for an asparagine residue. Infect Immun 1988; 56: 2144–8.
MICROBIAL PATHOGENESIS
Need for TolC, an Escherichia coli outer membrane protein, in the secretion of heat-stable enterotoxin I across the outer membrane Hiroyasu Yamanakaa, Tomohiko Nomuraa, Yoshio Fujiib & Keinosuke Okamoto∗c a
Department of Biochemistry, Faculty of Pharmaceutical Sciences, and bInstitute of Pharmacognosy, Tokushima Bunri University, Yamashiro, Tokushima, Tokushima 770-8514, Japan (Received April 13, 1998; accepted in revised form April 30, 1998)
Escherichia coli heat-stable enterotoxin Ip (STIp) is a typical extracellular toxin consisting of 18 amino acid residues synthesized as a precursor of pre (amino acid residues 1 to 19), pro (amino acid residues 20 to 54), and mature (amino acid residues 55 to 72) regions. STIp synthesized in the cytoplasm must cross the inner and outer membranes to migrate into the extracellular environment. Previous studies showed that the precursor translocates across the inner membrane utilizing the general export pathway consisting of Sec proteins. However, it remains unclear how it crosses the outer membrane. In this study, we examined the effects of mutation of the tolC gene which encodes an E. coli outer membrane protein, TolC, on the release of STIp into the extracellular environment. The mutation reduced the amount of STIp released into culture supernatant and increased the amount of STIp accumulated in the periplasm. This indicates that TolC mediates the translocation of STIp across the outer membrane. The inability to transfer STIp in the periplasm into the culture supernatant was restored by introduction of the tolC gene into the mutant cells. In the mouse intestinal loop assay, living cells of the mutants did not show a positive response, but wild-type cells did. These results showed that TolC is involved in the translocation of STIp 1998 Academic Press across the outer membrane. Key words: Escherichia coli, outer membrane, enterotoxin, TolC, secretion.
Introduction Exoproteins produced by gram-negative bacteria have to cross the cell envelope, which is ∗ Author for correspondence. 0882–4010/98/090111+10 $30.00/0
composed of two membranes (inner and outer membranes) separated by an aqueous periplasmic compartment containing the peptidoglycan cell wall [1, 2]. To date, three secretory pathways have been recognized in gram-negative bacteria; type I, type II and type III pathways [3–5]. Exoproteins employing the type I 1998 Academic Press
112
pathway are secreted across the bacterial envelope in a single step mediated by ABC (ATPbinding-cassette) transporter [6]. On the other hand, extracellular proteins secreted by the type II pathway are synthesized as precursor proteins with an amino-terminal signal sequence and are translocated across the inner membrane via the Sec protein machinery. The proteins released into the periplasm then cross the outer membrane mediated by an auxiliary secretion machinery composed of 13 to 15 accessory proteins [4, 7, 8]. The type III secretion system has been more recently identified and it has yet to be characterized in detail [3]. Escherichia coli heat-stable enterotoxin Ip (STIp) is an extracellular peptide toxin produced by enterotoxigenic E. coli strains [9]. The toxin is synthesized as a precursor consisting of three regions; pre (amino acid residues 1 to 19), pro (amino acid residues 20 to 54) and mature (amino acid residues 55 to 72) regions [10, 11]. Previous studies demonstrated that translocation of the STIp precursor across the inner membrane is achieved by the general export machinery consisting of Sec proteins, and the pre region of STIp functions as a signal peptide in translocation. After or during translocation across the inner membrane, the pre region is cleaved and the peptide generated, consisting of the pro and mature regions, is released into the periplasm [10]. Subsequently, we found that cleavage between the pro region and the mature region, and intramolecular disulfide bond formation in the mature region occur in the periplasmic space and that the mature STIp molecules thus generated cross the outer membrane and emerge in the culture supernatant [12, 13]. However, it remains unclear how the mature STIp molecules cross the outer membrane. As described above, the type II exoproteins released into the periplasmic space utilize the auxiliary secretion machinery composed of 13 to 15 accessory proteins to cross the outer membrane [4]. As the genes for these accessory proteins are not included in the normal E. coli chromosome, the type II exoproteins are not released into the culture supernatant of E. coli strains when only the protein gene (which does not contain the genes encoding the accessory proteins) are introduced into the cells [4]. However, STI is released into the culture supernatant from the E. coli transformed with plasmids encoding the STI gene but not any accessory protein genes [11]. Thus, the secretion machinery necessary for the translocation of STI across the
H. Yamanaka et al.
outer membrane is present in normal E. coli strains. The secretion therefore appears to be an intrinsic pathway in E. coli. It has been shown that TolC, an E. coli outer membrane protein [14], is involved in translocation of haemolysin [15] and colicin V [16] across the outer membrane. Furthermore, it has been reported that TolC is also related to the efflux of small molecules such as antibiotics [17, 18]. Therefore, we suspected that TolC is also involved in the translocation of STI across the outer membrane.
Results Disruption of the tolC gene of E. coli BL21 Strain BL21 has advantage in expression of recombinant proteins in E. coli, as the strain had lost both lon and ompT protease [19, 20]. Moreover, BL21 can be used as the host strain in the T7 expression system, as it contains T7 RNA polymerase gene on the chromosome [20]. We constructed a tolC-deficient mutant of BL21 by transduction with the P1 phage [21]. Escherischia coli CS1562 (k−, tolC6::Tn10) [22] was used as the parent strain for the mutation. After transduction with P1 phage, the cells grown in the presence of tetracycline (50 lg/ml) (Tn10 encodes the tetracycline resistance gene) were examined for sensitivity to SDS and colicin E1, characteristic properties accompanying tolC mutation as reported previously [23]. The cells that were sensitive to SDS and resistant to colicin E1 were selected as candidate tolC-deficient mutants. The tolC gene of the candidates was examined by PCR. The primers used were (i) and (ii). The sequence of the primer (i) is 5′ TTAAATGTCCTGGCACTAATA 3′ (sense) and that of (ii) is 5′ TGCGGCAGATAACCCGTATC 3′ (antisense). These extend from bases 101 to 120 and 1831 to 1850 of the tolC gene sequence reported by Niki et al. [24], respectively. A DNA fragment of the expected length (1750 bp) was amplified from the DNAs of wild-type strain (BL21), but no fragment was amplified from the DNAs of candidate cells (data not shown), suggesting that the tolC was mutated in these cells. To confirm the mutation of tolC, the production of the tolC mRNA from these strains was examined. Total cellular RNA was purified from the cells grown in L broth [25] at 37°C to
TolC and translocation of E. coli STI
exponential phase by the method described by MacDonald et al. [26]. The tolC mRNA in the sample was enzymatically amplified by reverse transcription reaction and PCR. The RT-PCR kit (R019A, Takara Shuzo Co., Kyoto, Japan) was used for the reaction. The sequences of the primers used in this experiment were (iii) 5′ TAACGCTGCAGGAAAAAGCA 3′ (sense) and (iv) 5′ TGCGGCAGATAACCCGTATC 3′ (antisense), extending from bases 618 to 637 and 1831 to 1850 of the tolC gene sequence reported by Niki et al. [24], respectively. The predicted length of DNA fragment amplified by the reaction is 1233 base pairs. The amplified DNAs were analysed by acrylamide gel electophoresis. The DNA fragment of 1233 base pairs was amplified from the sample of BL21, as expected, however, the fragment was not detected in the sample from the candidate cells (data not shown). This shows that tolC was mutated and was not transcribed in the candidate strain. The tolC mutant strain was named BL21-2.
Effect of tolC mutation on release of STI from cells To examine the effects of tolC mutation on the release of STI into culture supernatant, BL21 (tolC+) and BL21-2 (tolC−) were transformed with Tc-1 carrying STIp gene (Table 1). The transformants were cultured in L broth containing ampicillin (50 lg/ml) at 37°C with shaking for 5 h. The cells were separated from the culture supernant by centrifugation. The periplasmic fraction of the cells was prepared as described in Materials and methods and the volume of the fraction was adjusted to make it equal with that of the culture supernatant. The STI activity of these samples was measured using the suckling mouse assay. As shown in Fig. 1(a), the STI activity of the culture supernatant of BL21/Tc-1 was high, but that of BL21-2/Tc-1 was extremely low. In contrast, STI activity of the periplasmic fraction of BL21/Tc-1 was low and that of BL21-2/Tc-1 was high [Fig. 1(b)]. This indicated that STI produced in the tolC mutant cells did not translocate across the outer membrane of the mutant cells and accumulated in the periplasmic space.
Complementation of the tolC mutation To confirm the involvement of TolC in the translocation of STI, we examined whether the release
113
of STI into culture supernatant in the tolC mutant strain was restored by introducing the wild-type tolC gene into the tolC mutant. To carry out this experiment, Tc-1-TolC plasmid was constructed by inserting the tolC gene into Tc-1. As a control, the mutant tolC gene which does not produce functional TolC was inserted into Tc-1 as described in Materials and methods. The obtained plasmid carrying the mutant tolC gene was named Tc-1-mTolC (Table 1). The tolC mutant (BL21-2) was transformed with these plasmids and the transformed cells were cultured in L broth as described above. STI activity of the culture supernatant and of the periplasmic fraction of each strain was determined. The level of STI activity of culture supernatant of BL21-2/Tc-1-TolC was higher than that of the strain transformed with Tc-1 (BL21-2/Tc-1) and was close to that in wildtype cells (BL21/Tc-1) [Fig. 1(a)]. In contrast, the remaining STI activity in the cells of BL21-2/ Tc-1-TolC was very low [Fig. 1(b)]. These results indicated that the reduced efficiency of STI to cross the outer membrane caused by mutation of the tolC gene was restored by introducing the wild-type tolC gene into the mutant strain. However, the reduced efficiency was not restored by introducing the plasmid carrying the mutant tolC gene (Tc-1-mTolC) (Fig. 1). This further supports the notion that TolC is involved in the translocation of STI across the outer membrane.
Analysis by pulse-labelling The above results indicated that TolC is needed for STI to cross the outer membrane. However, the results were obtained by determination of STI activities of E. coli cells after long-term cultivation. It was possible that the STI activity of the mutant was affected by the side effects of the tolC mutation during long-term cultivation. Such side effects may lead to wrong conclusions. Analysis of the nascent STI may provide more accurate information regarding the effects of mutation of the tolC gene. We therefore analysed the nascent STI by pulse-labelling using pET system. pET11-STI plasmid contains STIp gene. pET11-STI-TolC plasmid and pET11-STI-mTolC plasmid were constructed by inserting the wild tolC gene and the mutant tolC gene (which does not produce functional TolC), respectively, into pET11-STI
114
H. Yamanaka et al.
Table 1. Strains and plasmids Strain or plasmid Strains E. coli BL21 CS1562 2086 BL21-2 Plasmids pBR322 Tc-1 pET11 pT7-Blue(R) pT7-TolC pT7-mTolC Tc-1-TolC Tc-1-mTolC pET11-STI pET11-STI-TolC pET11-STI-mTolC
Features
Source or reference
Containing T7 RNA polymerase gene on the chromosome E. coli K-12 derivative strain (F−, k−, tolC::Tn10) Clinical isolate producing STI TolC mutant of BL21
[20]
Vector plasmid Derivative of pBR322 containing STIp gene (parental plasmid for STIp gene) Containing a T7 promoter under the control of the lac operator Cloning vector for PCR-amplified gene Derivative of pT7-Blue(R) containing tolC gene Introduction of frame shift mutation at PstI site in the tolC gene of pT7-TolC Insertion of the tolC gene into BamHI site of Tc-1 Insertion of the mutant tolC gene encoded in pT7mTolC into BamHI site of Tc-1 Insertion of STIp structural gene of Tc-1 into NdeI site of pET11 under the control of the T7 promoter Insertion of tolC gene into HindIII-EcoRI site of pET11-STI Insertion of the mutant tolC gene encoded in pT7mTolC into HindIII-EcoRI site of pET11-Tc-1
(Table 1). BL21 and BL21-2 which were transformed with these pET11 derivatives were pulselabelled with l-[35S]cysteine for 3 min and chased for 3 min. BL21 cells harbouring pET11, which does not encode the STI gene, were used as a negative control. The volume of these cultures was 1 ml. The culture supernatant and the periplasmic fractions of the cultures were prepared as described in Materials and methods. The volume of each periplasmic fraction was adjusted to 0.4 ml. The samples prepared were resolved by SDS-PAGE. The samples from the culture supernatant and the periplasmic fraction were run in the gels and shown in Fig. 2(a) and (b), respectively. In the gel shown in Fig. 2(a), a dense band appeared in lanes 2 and 4 at a position corresponding to that of the mature STIp comprised of 18 amino acid residues [13]. This indicated that the culture supernatant from the wild-type strain (BL21/pET11-STI) contained STIp [lane 2, Fig. 2(a)], but that from the tolC mutant strain (BL21-2/pET11-STI) did not [lane 3, Fig. 2(a)].
[22] Our collection This study Takara Shuzo Co. [11] [20] Takara Shuzo Co. This study This study This study This study [13] This study This study
However, the culture supernatant from the mutant strain which was complemented with the native tolC gene (BL21-2/pET11-STI-TolC) contained STIp [lane 4, Fig. 2(a)]. Recovery of STI production was not achieved by introduction of the mutant tolC gene [lane 5, Fig. 2(a)]. Analysis of periplasmic samples showed that the density of the band corresponding to mature STIp (indicated by the arrow) was faint in the wild-type strain [BL21/pET11-STI, lane 2 of Fig. 2(b)]. This indicated that most mature STIp had already moved into the culture supernatant in the wild-type strain. However, the band in Fig. 2(b), lane 3, was dense. This sample was prepared from the periplasmic fraction of tolC mutant strain (BL21-2/pET11-STI). As shown in lane 3 of Fig. 2(a), the amount of STI released into culture supernatant from the mutant cells was very low. Thus, the STI which emerged into the periplasmic space of the mutant cells could not pass through the outer membrane. This accumulation of STI in the periplasmic space did not occur in the mutant strain complemented
TolC and translocation of E. coli STI
25
BL21/Tc-1
BL21-2/Tc-1
BL21-2/Tc-1-TolC
BL21/Tc-1-mTolC
BL21/Tc-1
BL21-2/Tc-1
BL21-2/Tc-1-TolC
BL21/Tc-1-mTolC
50
BL21/pBR322
Enterotoxic activity of tolC mutant
0
STI activity (unit)
the mutant tolC gene [lanes 5 of Figs 2(a) and (b)]. These results support the notion that TolC is directly involved in the translocation of STI across the outer membrane.
(a)
BL21/pBR322
STI activity (unit)
50
115
(b)
25
0
Figure 1. Effects of mutation of the tolC gene on STI activity of cells. BL21 (tolC-positive strain) and BL212 (tolC-negative strain) were transformed with plasmids Tc-1, Tc-1-TolC and Tc-1-mTolC carrying STIp, STIp and tolC, and STIp and the mutant tolC genes, respectively, as shown in Table 1. The transformed cells were cultured in L broth containing ampicillim (50 lg/ml) at 37°C for 5 h with shaking. After cultivation, the culture supernatants (a) and the periplasmic fractions (b) were prepared as described in the text. Enterotoxic activities of these samples were determined by the suckling mouse assay.
with the native tolC gene [BL21/pET11-STI-TolC, lane 4 of Fig. 2(b)] and most STI translocated across the outer membrane [lane 4, Fig. 2(a)]. However, the ability of the mutant cells to transport the STI in the periplasmic space into culture supernatant was not restored by introduction of
To clarify further the role of TolC in the translocation of STI across the outer membrane, the enterotoxic activity of tolC mutant cells carrying STIp gene was examined. BL21-2/Tc-1 and BL21/Tc-1 were each cultured on nutrient agar at 37°C for 6 h. The cells cultured were collected, washed with 0.5% NaCl and then suspended in L broth at a concentration of 1.5×107 colonyforming units (cfu)/ml. A portion of the cell suspension (0.2 ml) was injected into a mouse intestinal loop. Three hours later, the amount of fluid accumulated in the loop was determined. As shown in Fig. 3, BL21/Tc-1 induced a marked positive response {fluid accumulation ratio [weight of loop (g)/length of loop (cm)] above 0.20 was regarded as a positive response}, while BL21-2/Tc-1 did not. This indicated that sufficient amounts of STI for accumulation of the fluid in the intestinal lumen were not released from the tolC-negative strain (BL21-2/Tc-1), although the amount of STI released from the tolC-positive strain (BL21/Tc-1) was sufficient. These observations indicated that TolC is important for release of STI not only in cells cultured in vitro but also in those inoculated into the intestinal lumen of mice.
Discussion We analysed the translocation of STI across the outer membrane of E. coli and found that TolC is an indispensable element in transport of STI from the periplasmic space to the culture supernatant in BL21. However, it remained unclear whether TolC functions in clinical isolates as well as in the strains using for gene manipulation. As described above, BL21 had lost both lon and ompT proteases. The decomposition rate of proteins synthesized is therefore thought to be slower in BL21 compared with clinical isolates. Furthermore, the plasmid encoding the STI gene used in this experiment was a multicopy plasmid. Thus, the amounts of STI synthesized in strains harbouring the plasmid would be
1
2
3
4
BL21-2/pET11-STI-mToIC
BL21-2/pET11-STI-TolC
(b)
5
5
BL21-2/pET11-STI-mToIC
BL21-2/pET11-STI
4
BL21-2/pET11-STI-TolC
3
BL21-2/pET11-STI
2
BL21/pET11-STI
1
BL21/pET11
(a)
BL21/pET11-STI
H. Yamanaka et al.
BL21/pET11
116
higher than that in clinical isolates. Therefore, the STI secretion pathway in the strain used in this study may have been different from that of clinical isolates, and therefore the conclusions guided from the results of this experiment could only be applied to the BL21 strain harbouring multicopy plasmid. To confirm the role of TolC in clinical isolates, we mutated the tolC gene of several clinical isolates of E. coli producing STI by P1 transduction and examined the amounts of STI released into the culture supernatant of the mutant cells. The amount was low in all mutants (data not shown), showing that TolC is important for the release of STI into the culture supernatant not only in BL21 but also in the clinical isolates. Therefore, it can be concluded that TolC is needed for STI to cross the outer membrane in all E. coli strains. TolC has been reported to be involved in the secretion of both haemolysin and colicin V [15, 16]. These proteins pass through the inner and outer membranes in an all-or-nothing fashion; that is, the proteins synthesized in the cytoplasm utilize a secretory machinery consisting of a translocator in the inner membrane and of TolC in the outer membrane. The inner membrane translocator is comprised of an ABC exporter and an accessory protein [27]. It is said that the accessory protein acts as a bridge to interact with TolC in outer membrane [18]. As the components comprising the machinery are connected, proteins which utilize the secretory machinery translocate across the two
Figure 2. Effects of mutation of the tolC gene on translocation of STI synthesized in cells across the outer membrane. To selectively express STI, the T7 expression system was used. BL21 (tolC-positive strain) and BL21-2 (tolC-negative strain) were used as host strains. These host strains were transformed with derivatives of pET11 (Table 1). The transformed cells were cultured in L broth until reaching the exponential phase. After induction with 2 mM IPTG, the cells were labelled for 3 min with l-[35S]cysteine (10 lCi). The labelling reaction was terminated by adding cold cysteine into the cultures, and incubation was continued for 3 min (chase period). After the chase period, the culture supernatants and the periplasmic fraction of cells were prepared as described in the text. The samples were resolved by SDSPAGE and radioactive bands were detected by image analysis. The arrow indicates the position of the mature STIp. (a) culture supernatant; (b) periplasmic fraction.
TolC and translocation of E. coli STI
117
Postive response
0.2
Negative response
Fluid accumulation (g/cm)
0.3
BL21-2/Tc-1
BL21/Tc-1
0.1
Cells inoculated
Figure 3. Enterotoxic activities of tolC mutants. BL21 (tolC-positive strain) and BL21-2 (tolC-negative strain) were transformed with Tc-1 carrying the STIp gene. The cells which were grown on nutrient agar were suspended in L broth at a concentration of 1.5×107 cfu/ml. A portion of the cell suspension was injected into a mouse intestinal loop. Three hours later, the amount of fluid accumulated in the loop was determined. Fluid accumulation ratio [weight of loop (g)/length of loop (cm)] above 0.20 was regarded as a positive response.
membranes without stopping. Therefore, the protein either crosses both membranes or remains in the cytoplasm, with no intermediate protein detected in the periplasm. In contrast, STI translocates across the inner membrane in a SecA-dependent manner, and intermediates of STI have been detected in the periplasm [12, 13]. The results of the present study showed that TolC is involved in the movement of STI from the periplasm to the extracellular environment. As TolC in the outer membrane forms a trimer [14], STI might pass through the pore constructed by association of three molecules of TolC.
As described above, TolC is regarded as a component of the secretory machinery and is thought to be involved in the translocation of protein through both inner and outer membranes in an all-or-nothing fashion. The present study showed that TolC also functions in the translocation of proteins brought into the periplasmic space by Sec proteins across the outer membrane. From these results, it can be concluded that TolC functions in translocation across the outer membrane not only of proteins which pass through the two membranes in an all-or-nothing fashion but also of proteins which are carried into the periplasmic space in a SecAdependent manner. Such a notion about TolC was first proposed by Foreman et al. [28]. They examined the secretion of STII, another heat-stable enterotoxin of E. coli, in which the amino acid sequence is quite different from that of STI, into culture supernatant and found that the efficiency of the release of STII into culture supernatant was low in a tolC mutant strain. As the translocation of STII across the inner membrane is achieved by the machinery consisting of TolC from the result. Recently, we also examined the effect of the mutation of the tolC gene on the secretion of STII into culture supernatant by the system used in this study. The result also showed that the efficiency of translocation of STII across the outer membrane was reduced by the tolC mutation. However, a small amount of STII was released from the mutant cells (unpublished data). This indicates that the mutation of tolC affects the secretion of STII. However, it is not elucidated that TolC is directly involved in the translocation of STII across the outer membrane. Further studies are necessary to clarify the role of TolC in the secretion of STII. The mechanism by which TolC recognizes STI in the periplasm might be quite different from those involved in the recognition of the haemolysin and colicin V. Recent studies showed that TolC is also involved in the efflux of small toxic components including antibiotics [17]. Therefore, studies of the recognition of STI by TolC might be useful to gain insight into the characteristics of the antibiotic-resistant bacteria.
Materials and methods Bacterial strains and plasmids The E. coli strains used in this study are listed in Table 1. E. coli CS1562 (k−, tolC6::Tn10) [22],
118
kindly provided by B. J. Backmann (E. coli Genetic Stock Center, Yale University, New Heaven, Conn., U.S.A.), was used as the tolC mutant parental strain. Plasmid Tc-1 encoding the STIp gene was the wild-type plasmid [11]. To selectively express STIp, the pET system was used [20]; in pET11STI, the STIp gene was cloned into NdeI of pET11 bringing its production under control of the T7 promoter [13].
Cloning and mutation of tolC gene The DNA fragment of the tolC gene was cloned as follows. The tolC gene was amplified from the chromosomal DNA of an STI-producing E. coli clinical isolate by PCR (strain 2086, Table 1). The primers used are (i) and (ii) which are described in Results. The DNA fragments amplified with these primers encompassed both the promoter and the structural regions of the tolC gene. The fragments were ligated into pT7Blue(R) (Takara Shuzo Co.), a cloning vector for PCR-amplified DNA fragments. The plasmid obtained was named pT7-tolC (Table 1). The mutant tolC gene was prepared as follows, pT7-TolC has two PstI sites, one in the tolC gene and the other in the multi-cloning site of pT7Blue(R). The nucleotide sequence at the position which is recognized by PstI in the tolC gene encodes the amino acid residue at position 107 from the amino terminus of TolC [24]. As TolC is composed of 495 amino acid residues, this frame shift mutation at the PstI site results in production of a mutant protein which is nonfunctional. To introduce the frame shift mutation, pT7-TolC was partially digested with PstI, and the resultant linear fragment was purified. The fragment was treated with Klenow to digest the protruding terminus at 3′ single-stranded regions, and the fragments generated were ligated. Then, the plasmid containing a frame shift mutation at the PstI site in the tolC gene was selected. The plasmid obtained was designated pT7-mTolC (Table 1).
Insertion of tolC into plasmids carrying the STI gene Insertion of the tolC gene into Tc-1 was carried out as follows: pT7-TolC was digested with HindIII and BamHI and the resulting linear fragment containing tolC gene was isolated. The
H. Yamanaka et al.
fragment was treated with Klenow to fill in the 5′ single-stranded ends. The fragment was cloned into the Klenow-treated BamHI site of Tc-1. The plasmid obtained was designated Tc1-TolC. Similarly, the mutant tolC gene isolated from pT7-mTolC was inserted into Tc-1, and the plasmid obtained was designated Tc-1-mTolC (Table 1). Insertion of the tolC gene into pET11-STI was carried out as follows: pT7-TolC was digested with HindIII and EcoRI and the resulting linear fragment containing tolC gene was isolated. The fragment was ligated to the HindIII-EcoRI fragment of pET11-STI encoding the STIp gene. The plasmid obtained was designated pET11-STITolC. Similarly, the mutant tolC gene isolated from pT7-mTolC was inserted into pET11-STI, and the plasmid obtained was designated pET11-STI-mTolC (Table 1).
Pulse-labelling BL21 cells transformed with pET11 derivative plasmids were shaken at 37°C until they reached an optical density at 660 nm of 0.4 in L broth containing ampicillin (50 lg/ml). The cultures were centrifuged, and the cell pellets were suspended in M9 minimal medium [25] containing an amino acid mixture (final concentration of each amino acid, 50 lg/ml) without cysteine at an optical density at 660 nm of 3.0. Afer adding a final concentration of 2 mM isopropyl-b-dthiogalactopyranoside (IPTG), the suspension was shaken at 37°C for 20 min. A portion (1 ml) was labelled for 3 min at 37°C with 10 lCi of l[35S]cysteine (0.5 to 1 kCi/mmol, ARS-101: American Radiolabeled Chemicals, Inc., St Louis, MO, U.S.A.) per ml. Incorporation of the label was terminated by adding 250 lg of unlabelled cysteine. The incubation was continued for 3 min. Thereafter, the supernatant was separated from the cells by centrifugation. The periplasmic fraction of the cells was prepared as described below. Prepared samples were mixed with SDS dye solution [29], heated at 100°C for 10 min and resolved by SDS-20% polyacrylamide gel electrophoresis (PAGE) [29]. The radioactive bands were visualized by autoradiography using an imaging plate analyzer (BAS 2000, Fujix, Tokyo, Japan).
Preparation of the periplasmic fraction Cells cultured in liquid medium were collected by centrifugation at 12 000 g for 10 min, then the
TolC and translocation of E. coli STI
cell pellet was suspended in 10 mM Tris-HCl buffer (pH 7.5). To prepare the periplasmic fraction, the cell suspension was incubated with polymyxin B at a concentration of 6500 U/mol at 4°C for 15 min and centrifuged (12 000 g for 15 min). The supernatant obtained was used as the periplasmic fraction.
Assay of STI activity STI activity was determined by means the suckling mouse assay as described [30]. Briefly, 0.1ml samples were administered via a gastric tube into the stomach of 2–3-day-old suckling mice, using 0.01% Evans blue due as a marker. The mice were killed 3.0 h later, and a ratio of intestinal to body weight of over 0.083 was considered positive. One unit was defined as the minimum effective dose that elicited a positive response, and the enterotoxin titre of samples was expressed as the reciprocal of the highest dilution that gave 1 unit of enterotoxin activity per 0.1 ml. The enterotoxic activity of each sample was determined in five mice.
Mouse intestinal loop assay Mice weighing 30–35 g were anaesthetized with sodium pentobarbital, and the intestines were exteriorized through a midline incision. The intestinal lumen was rinsed three times with saline. After rinsing, a series of ligated intestinal segments (loops), about 3 cm long and separated by a 0.5 to 1 cm interloop, were created. The most proximal loop was placed about 3 cm distal to the ligament of Treitz. One or two loops were created per intestine, and each loop was injected with 0.2 ml of sample solution. Three hours after injection, the mice were killed and the sample activities were measured and expressed as the ratio of the weight of the loop (in g) to its length (in cm). A ratio of over 0.20 was regarded as a positive response.
Acknowledgement This work was supported in part by a grant-in-aid for scientific research from the Ministry of Education, Science and Culture, Japan.
119
References 1 Nikaido H. Outer membrane. In: Neidhardt FC, Curtiss III R, Ingraham JL, Lin ECC, Low Jr. KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE, Eds. Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd edn. Waashington, D.C.: American Society for Microbiology, 1996; 29–47. 2 Wandersman C. Secretion across the bacterial outer membrane. In: Neidhardt FC, Curtiss III R, Ingraham JL, Lin ECC, Low Jr. KB, Magasanik B, Reznikoff WS, Riley M, Schaechter M, Umbarger HE, Eds. Escherichia coli and Salmonella: Cellular and Molecular Biology, 2nd edn. Washington, D.C.: American Society for Microbiology, 1996; 955–966. 3 Lee CA. Type III secretion systems: machines to deliver bacterial proteins into eukaryotic cells? Trends Microbiol 1997; 5: 148–56. 4 Pugsley AP. The complete general secretory pathway in Gram-negative bacteria. Microbiol Rev 1993; 57: 50–108. 5 Salmond GPC, Reeves PJ. Membrane traffic wardens and protein secretion in gram-negative bacteria. TIBS 1993; 18: 7–12. 6 Fath MJ, Kolter P. ABC transporters: bacterial exporters. Microbiol Rev 1993; 57: 995–1017. 7 Arkowitz RA, Bassilana M. Protein translocation in Escherichia coli. Biochim Biophys Acta 1994; 1197: 311–43. 8 Den Blaauwen T, Driessen AJM. Sec-dependent preprotein translocation in bacteria. Arch Microbiol 1996; 165: 1–8. 9 Takao T, Hitouji T, Aimoto S, et al. Amino acid sequence of a heat-stable enterotoxin from enterotoxigenic Escherichia coli 18D. FEBS Lett 1983; 152: 1–5. 10 Okamoto K, Takahara M. Synthesis of Escherichia coli heat-stable enterotoxin STp as a pre-pro form and role of the pro sequence in secretion. J Bact 1990; 172: 5260–5. 11 So M, McCarthy BJ. Nucleotide sequence of bacterial transposon Tn 1681 encoding a heat-stable (ST) toxin and its identification in enterotoxigenic Escherichia coli strains. Proc Natl Acad Sci USA 1980; 77: 4011–15. 12 Yamanaka H, Kameyama M, Baba T, Fujii Y, Okamoto K. Maturation pathway of Escherichia coli heat-stable enterotoxin I: requirement of DsbA for disulfide bond formation. J Bact 1994; 176: 2906–13. 13 Yamanaka H, Nomura T, Fujii Y, Okamoto K. Extracellular secretion of Escherichia coli heat-stable enterotoxin I across the outer membrane. J Bact 1997; 179: 3383–90. 14 Koronakis V, Li J, Koronakis E, Stauffer K. Structure of TolC, the outer membrane component of the bacterial type I efflux system, derived from two-dimensional crystals. Mol Microbiol 1997; 23: 617–26. 15 Wandersman C, Delepelaire P. TolC, an Escherichia coli outer membrane protein required for hemolysin secretion. Proc Natl Acad Sci USA 1990; 87: 4776–80. 16 Gilson L, Mahanty HK, Kolter R. Genetic analysis of an MDR-like export system: the secretion of colicin V. EMBO J 1990; 9: 3875–84. 17 Fralick JA. Evidence that TolC is required for functioning of the Mar/AcrAB efflux pump of Escherichia coli. J Bact 1996; 178: 5803–5. 18 Paulsen IT, Park JH, Choi PS, Saier M. A family of gram-negative bacterial outer membrane factors that function in the export of proteins, carbohydrates, drugs
120
19 20 21 22
23
24
and heavy metals from gram-negative bacteria. FEMS Microbiol Lett 1997; 156: 1–8. Phillips TA, VanBogelen RA, Neidhardt FC. lon gene product of Escherichia coli is a heat-shock protein. J Bact 1984; 159: 283–7. Studier FW, Rosenberg AH, Dunn JJ, Dubendorff JW. Use of T7 RNA polymerase to direct expression of cloned genes. Methods Enzymol 1990; 185: 60–89. Silhavy TJ, Berman ML, Enquist LW. Experiments with Gene Fusions. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory, 1984. Austin EA, Graves JF, Hite LA, Parker CT, Schnaitman CA. Genetic analysis of lipopolysaccharide core biosynthesis by Escherichia coli K-12: insertion mutagenesis of the rfa locus. J Bact 1990; 172: 5312–25. Binet R, Wandersman C. Cloning of the Serratia marcescens hasF gene encoding the has ABC exporter outer membrane component: a TolC analogue. Mol Microbiol 1996; 22: 265–73. Niki H, Imamura R, Ogura T, Hiraga S. Nucleotide
H. Yamanaka et al.
25 26 27 28
29 30
sequence of the tolC gene of Escherichia coli. Nucl Acids Res 1990; 18: 5547. Miller JH. Experiments in Molecular Genetics. Cold Spring Harbor, N.Y.: Cold Spring Harbor Laboratory, 1972. MacDonald RJ, Swift GH, Przybyla AE, Chirgwin JM. Isolation of RNA using guanidinium salts. Methods Enzymol 1987; 152: 219–27. Higgins CF. ABC transporters: from microorganisms to man. Ann Rev Cell Biol 1992; 8: 67–113. Foreman D, Martinez I, Coombs G, Torres A, Kupersztoch YM. TolC and DsbA are needed for the secretion of STb, a heat-stable enterotoxin of Escherichia coli. Mol Microbiol 1995; 18: 237–45. Laemmli UK. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970; 227: 680–5. Okamoto K, Okamoto K, Yukitake J, Miyama A. Reduction of enterotoxic activity of Escherichia coli heatstable enterotoxin by substitution for an asparagine residue. Infect Immun 1988; 56: 2144–8.