Identification of a protein secretory pathway for the secretion of heat-labile enterotoxin by an enterotoxigenic strain of Escherichia coli

Bacterial Strains, Plasmids, and Growth Conditions.

The bacterial strains and plasmids used in this study are listed in Table 1 (21, 22). Strains were routinely grown at 37°C in Luria broth (1% tryptone/0.5% yeast extract/171 mM NaCl) (23) or CAYE medium [2% Casamino acids/0.6% yeast extract/43 mM NaCl/38 mM K₂HPO₄/0.25% glucose/0.1% (vol/vol) trace salts solution consisting of 203 mM MgSO₄/25 mM MnCl₂/18 mM FeCl₃] (24) supplemented with 100 μg/ml of ampicillin (CSL, Parkville, Victoria, Australia) or 50 μg/ml of kanamycin (Sigma–Aldrich) when necessary. Y1 mouse adrenal cells (CSL) were maintained in DMEM (23) supplemented with 10% FBS (CSL) and 100 μg/ml of gentamicin (Sigma–Aldrich) and incubated at 37°C in an atmosphere containing 5% CO₂.

Table 1.

Bacterial strains and plasmids used in this study

Bacterial strain or plasmid	Relevant characteristics	Ref. or source
E. coli
H10407	ETEC serotype O78:H11, LT+ ST+	21
MT13	H10407gspD∷KmR	This study
XL1-Blue	F′∷Tn10 proA+B+lacIq Δ(lacZ)M15/recA1 endA1 gyrA96 (NalR) thi hsdR17(rm)supE44 relA1 lac	Stratagene
Plasmids
pST76-A	Temperature-sensitive suicide vector derived from pSC101, ApR	22
pUC4-KIXX	pUC4K derivative containing KmR and BmR genes from Tn5	Pharmacia
pGEM-T Easy	High copy number vector, ApR	Promega
pMT37	pGEM-T Easy containing 1.9-kb PCR fragment encompassing gspD from H10407, ApR	This study
pMT39	pMT37 containing 1.4-kb SmaI fragment encompassing the KmR gene from pUC4-KIXX cloned into SnaBI site, ApR KmR	This study
pMT42	pST76-A containing 3.3-kb EcoRI fragment of pMT39 bearing gspD with an inserted KmR gene cloned into the EcoRI site, ApR KmR	This study
pMT6	pGEM-T Easy containing 5.5-kb PCR fragment encompassing gspD, gspE, and gspF from H10407, ApR	This study
pMT44	Deletion derivative of pMT6 containing 3.8-kb fragment encompassing gspD from H10407, ApR	This study

Open in a new tab

Ap^R, ampicillin resistance; Km^R, kanamycin resistance; Bm^R, bleomycin resistance. 

Growth and Viability Studies.

Bacterial growth was assayed by measuring the optical density at OD₆₀₀ of cultures every hour by using an Ultrospec III spectrophotometer (Pharmacia). To determine the extent of cell death (lysis), the Molecular Probes Baclight Live/Dead kit was used (Molecular Probes). Both live and dead bacteria were labeled simultaneously as instructed by the manufacturer, after which they were visualized by using a Leica DM-LB HC microscope with a I3 fluorescein filter (excitation 450–490 nm). The percentage of live and dead bacteria was ascertained by counting at least 100 cells in at least five separate random fields.

Preparation and Manipulation of DNA.

Plasmid DNA was isolated by using the Wizard Plus SV DNA Purification System (Promega). Standard restriction digestion and cloning procedures using DNA-modifying enzymes supplied by Promega or New England Biolabs were used (23). Electrocompetent E. coli XL1-Blue, H10407, and MT13 cells were obtained by growing the bacteria to midlogarithmic phase (OD₆₀₀ 0.5–0.8) at 37°C. Cells were then washed three times with sterile cold 10% glycerol in distilled water and resuspended in 1/75 the original culture volume. Transformation was achieved by using a Bio-Rad Gene Pulser and electroporation conditions 1.80 kV, 200 Ω, and 25 μF.

DNA Amplification, Sequencing, and Analysis.

PCR was used to amplify the genes for the putative type II secretion pathway and to confirm insertional inactivation of gspD by allelic exchange. Genomic DNA of E. coli H10407 was prepared by the boiling lysis method (23). PCR amplifications were performed with Vent DNA polymerase (New England Biolabs), which has proofreading activity, or with AmpliTaq polymerase (Applied Biosystems) in a reaction volume of 25–50 μl in a Gene Amp PCR System 9700 thermal cycler (Applied Biosystems). The PCR conditions involved denaturation for 5 min at 94°C, followed by 35 cycles of 30 sec at 94°C, 30 sec at 55°C, and 1–4 min at 72°C, with a final extension for 5 min at 72°C.

Nucleotide sequencing was performed by using an abi prism Big Dye Terminators, Ver. 3.0, Cycle Sequencing Kit (Applied Biosystems). Reactions were analyzed on an Applied Biosystems ABI PRISM 377 DNA sequencer. Nucleotide sequence data were edited and assembled into contiguous sequence with the sequencher program (Gene Codes, Ann Arbor, MI).

Nucleotide sequence and amino acid similarity searches with sequences in the public databases were performed by using the blastn, blastp, and blastx programs (25) available from the National Center for Biotechnology Information web site (www.ncbi.nlm.nih.gov).

Construction of a Functionally Nonpolar gspD Mutant.

A 1.9-kb fragment containing 1.6 kb of gspD was amplified by PCR from E. coli H10407 genomic DNA by using oligonucleotide primers P35 (5′-ATTGTCGGTTATGCAGCGAAGC) and P109 (5′-TCCACCTTCGAGACTTCC). The PCR product was ligated into pGEM-T Easy and electroporated into XL1-Blue to generate pMT37. A SmaI flanked kanamycin-resistance gene was excised from pUC4-KIXX and inserted into the unique SnaBI site of pMT37 generating plasmid, pMT39. A 3.3-kb EcoRI fragment of pMT39, which contained the kanamycin-resistance gene and flanking gspD sequences, was ligated into the EcoRI site of the temperature sensitive suicide vector pST76-A (22). The resulting construct, pMT42, was transformed into E. coli H10407 by electroporation. Bacteria were plated on LB agar containing kanamycin and incubated at 30°C overnight. Transformants of E. coli H10407 were then incubated at 42°C for 24 h to obtain cells cured of the free plasmid. Bacterial colonies were replica plated on LB agar containing ampicillin or kanamycin to identify ampicillin-sensitive kanamycin-resistant transformants. One of these was E. coli MT13, in which inactivation of gspD by insertion of the kanamycin-resistant gene was confirmed by using PCR analysis.

Complementation of the gspD Mutant.

A 5.5-kb fragment containing gspD, gspE, and gspF was amplified by PCR from E. coli H10407 genomic DNA by using Vent DNA polymerase and oligonucleotide primers P35 (5′-ATTGTCGGTTATGCAGCGAAGC) and P36 (5′-ACTCCTGCAAATTCCAGTTACC). The PCR product was A-tailed with AmpliTaq polymerase, ligated into pGEM-T Easy behind the lac promoter, and electroporated into XL1-Blue to generate pMT6. The correct orientation of the insert was verified by restriction analysis, and its sequence was determined to ensure that fidelity had been maintained during amplification. pMT6 was digested with NcoI to generate a deletion derivative, pMT44, in which only gspD was retained on a 3.8-kb fragment. pMT44 was then transformed into E. coli MT13 by electroporation.

Preparation of Culture Supernatants and Periplasmic Extracts.

E. coli H10407, MT13, and MT13(pMT44) were cultured overnight in 5 ml of LB broth at 37°C in a rotary shaker set at 180 rpm. Five milliliters of prewarmed CAYE medium in 50-ml Erlenmeyer flasks was then inoculated with 125 μl of the overnight bacterial cultures and incubated at 37°C with vigorous aeration (220 rpm in a rotary shaker). To isolate culture supernatants, 4-, 8-, and 24-h cultures were clarified of cells by centrifugation at 3,000 × g for 20 min at 4°C and by passage through a filter with a 0.20-μm-diameter pore size (Sartorius).

To prepare periplasmic extracts, bacterial pellets from cultures used to prepare the supernatants were washed in 1 ml of PBS, pH 7.2 and resuspended in an equal volume of PBS containing 12,000 units/ml polymyxin B (Sigma–Aldrich). After incubation with gentle shaking at 37°C for 30 min, the periplasmic extracts were separated from cell debris by centrifugation at 16,000 × g for 3 min. They were then filtered as described above, after which the volume of the extract was adjusted to match that of the culture supernatant. Supernatants and periplasmic extracts were assayed immediately or stored at 4°C and assayed within 48 h.

The integrity of the supernatant and periplasmic fractions was determined by assaying these fractions for the periplasmic enzyme alkaline phosphatase by using the Sigma 104 alkaline phosphatase kit as follows: 0.05 ml of 221 phosphate buffer, 0.05 ml of p-nitrophenyl phosphate (PNPP), and 0.01 ml of either periplasmic or supernatant extracts were incubated at room temperature for 30 min, after which the reactions were stopped by the addition of 0.1 ml of 5.0 M NaOH. Hydrolysis of PNPP was detected at OD₄₂₀.

Assays for LT.

LT was assayed for receptor binding by enzyme immunoassay (EIA) (26) and for biological activity by using Y-1 adrenal cells (27). The EIA was modified from a previously described method (26), with all washes performed three times by using PBS containing 0.5% (vol/vol) Tween-20 (PBS-T), and all samples, standards, and antibodies diluted in PBS-T containing 1% (vol/vol) FBS. In addition, all incubations were at 37°C. Briefly, microtiter plates (Nalge Nunc) were coated with 0.1 μg of G_M1 ganglioside (Sigma–Aldrich) in 0.06 M sodium carbonate–bicarbonate buffer (pH 9.6) and incubated overnight at 4°C. Plates were washed, blocked with 5% (vol/vol) FBS in PBS for 1 h, and washed again. After initial dilution (culture supernatants 1 in 2 and periplasmic extracts 1 in 5), samples were serially diluted 5-fold and added to the wells. Standard curves were generated in each assay plate by using 2-fold serial dilutions of purified LT (Sigma–Aldrich), at a starting concentration of 400 ng/ml. Plates were incubated for 2 h, washed, and incubated for 1 h with rabbit anti-CT antiserum (Sigma–Aldrich) diluted 1 in 5,000. After washing, horseradish peroxidase-conjugated sheep anti-rabbit IgG (Sigma–Aldrich), diluted 1 in 1,000, was added and incubated for 1 h. After further washing, bound antibody was detected by using the chromogenic substrate, TMB (Kirkegaard & Perry Laboratories). After 3 min at room temperature, the reaction was stopped by the addition of 2 M H₂SO₄ (30 μl/well), and the OD₄₅₀ was measured in an ELX800 Universal microplate reader (Bio-Tek, Winooski, VT). Each assay was performed in triplicate on at least three separate occasions. Regression analysis (R² > 0.96) was used to generate a standard curve for determination of LT concentrations in the test samples.

To measure biological activity of the toxin preparations, Y1 mouse adrenal cells were seeded in 96-well microtiter plates at a concentration of 2 × 10⁴ cells per well. After 48 h, the growth medium was replaced with DMEM supplemented with 1% FBS and gentamicin (100 μl), containing duplicate samples of bacterial culture supernatant and periplasmic extracts, serially diluted 2-fold after an initial dilution of 1 in 10. The cells were incubated for 18 h and then examined for typical rounding by a blinded operator by using phase-contrast microscopy at a magnification of ×200. The end point of the assay was defined as the highest dilution that showed more than 50% cell rounding. Each assay was performed in duplicate on at least two separate occasions. Two-fold dilutions of purified LT at a starting concentration of 10 ng/ml were assayed in the same plates and the result used to determine the sensitivity of the assay.

Prevalence of the Type II Secretion Pathway in ETEC Strains.

To determine the prevalence of the novel type II secretion pathway in ETEC strains of clinical origin, a variety of ETEC isolates of different serotypes and toxin profiles were examined by PCR for a 1.0-kb fragment of gspD by using primers P78 (5′-TTCGGAAATCGCCCGCGTGC) and P109 (5′-TCCACCTTCGAGACTTCC), and for a 1.2-kb fragment of gspK by using primers P5 (5′-GCAGCAGGTGACTAACGGC) and P12 (5′-CAGGGCTTAACCACGGGTC).

Statistical Analysis.

All analyses were performed by using Student's t test. A value of P < 0.05 was taken to indicate statistical significance.

Nucleotide Sequence Accession Number.

The E. coli H10407 type II secretion locus has been assigned GenBank accession no. AY056599.

Identification, Sequencing, and Characterization of the Type II Secretion Genes.

In our work, to characterize pathogenicity islands in rabbit-specific strains of enteropathogenic E. coli (28), we identified and sequenced a novel type II protein secretion pathway from one particular strain (GenBank accession no. AF426313). Using primers generated from this sequence, we found that several other pathogenic strains of E. coli possessed this pathway, including the prototypic human ETEC strain H10407. We then determined the sequence of the H10407 chromosome encompassing the putative genes for this pathway by amplifying and sequencing overlapping fragments (GenBank accession no. AY056599). Computer analysis of 11,765 bp of this sequence revealed 13 ORFs, 11 of which were organized in a single operon, with several of the ORFs overlapping. Because the genetic organization of the type II gene cluster is relatively well conserved (29), we designated the newly identified genes gspC-M to comply with current nomenclature (Fig. 1A). Interestingly, a search of the public databases for homologous sequences revealed that the 5′ and 3′ ends of the ETEC type II secretion pathway are also present in the E. coli K-12 strain, MG1655 (15) (Fig. 1B).

The first complete ORF of the ETEC type II secretion gene cluster is pppA; PppA is 96% identical to its homologue in MG1655. The next closest homology (46% identity, 63% similarity) is to VcpD (PilD), the prepilin peptidase required for the secretion of CT by V. cholerae (30, 31). Downstream of pppA is a small ORF, yghG, which encodes a putative outer membrane lipoprotein, showing 24% identity and 49% similarity to a conserved hypothetical protein of unknown function from V. cholerae, but which is not colocated with either vcpD (pilD) or the type II gene cluster. The proteins encoded by the genes gspC-M, which evidently comprise a single operon, show the closest homology to the type II secretion proteins of V. cholerae (10), with sequence identities ranging from 76 to 24% and similarities ranging from 86 to 47% (Fig. 1C).

To determine the prevalence of the type II secretion pathway in clinical isolates of ETEC, we performed PCR analysis for gspD and gspK on a variety of ETEC strains of serotypes O6:H16, O8:H9, O20:H-, O25:H-, O25:H42, O75:H4, O78:H11, O114:H21, O128:H21, O148:H28, O149:H10, O27:H20, O159:H34, and O-nontypeable:H9. The test bacteria included 11 that produced LT only, 9 that produced both LT and heat-stable enterotoxin, and 4 that produced heat-stable enterotoxin only. All 24 of these strains tested positive for both gspD and gspK.

Determination of the Role of the Type II Pathway in the Secretion of LT by ETEC Strain H10407.

Because V. cholerae requires a type II protein secretory pathway to secrete CT, we investigated whether the homologous pathway of E. coli H10407 played a role in the secretion of LT. To this end, we inactivated the pathway by inserting a kanamycin resistance gene into gspD by homologous recombination (Fig. 1A). We chose this gene for inactivation because in Pseudomonas aeruginosa, the homologue of GspD forms multimeric complexes, is localized within the outer membrane, and is the putative pore of the type II secretion apparatus (32, 33). Insertional inactivation of gspD in ETEC H10407 yielded the mutant strain, MT13.

To determine whether MT13 was able to secrete LT, we used an EIA (26) to compare LT secretion by MT13 with that by E. coli H10407. Comparison of the growth curves of E. coli H10407 and MT13 showed no difference between them and indicated that suitable time points to assay the culture medium for LT were 4 h (representing midlogarithmic phase), 8 h (early-stationary phase), and 24 h (late-stationary phase) (Fig. 2A). The detection limit of the EIA using purified LT was determined to be 0.4 ng/ml, with a standard curve that was linear between 2.6 and 24.0 ng/ml (data not shown). This portion of the curve was used to estimate the amount of LT in the test samples. The results of these assays (Fig. 2B) showed that by 4 h, H10407 had secreted 58.5 ± 3.7 (mean ± SD) ng/ml of LT into the supernatant, but the amount of LT in the supernatant of the gspD mutant MT13 was below the level of detection (<0.4 ng/ml; P = 0.001). At the same time point, however, the amount of LT in the periplasm of the mutant (31.3 ± 2.6 ng/ml) was approximately 10-fold greater than that of the wild type (3.4 ± 0.1 ng/ml; P < 0.001) (Fig. 2C). These results indicate that MT13 could synthesize LT but could not secrete it. By 8 h the amount of LT secreted into the medium by H10407 had increased to 90.3 ± 7.6 ng/ml, but the amount of LT in the supernatant of MT13 was still below the level of detection (P = 0.002). At the same time point the amount of toxin retained in the periplasm of the mutant had fallen to 1.5 ± 0.1 ng/ml, but was still significantly greater than that found in the wild type, which was below detectable levels (P = 0.002). The reduction in the amount of LT in the periplasm of the mutant could be due to proteolytic degradation/turnover (14), as the A subunit is susceptible to a variety of proteases (11). By 24 h, E. coli H10407 had secreted 193.6 ± 1.2 ng/ml of LT into the medium. At this time point, 29.1 ± 1.1 ng/ml of LT was detectable in the supernatant for MT13 (P < 0.001), but the amount of toxin retained in the periplasm of MT13 was still significantly greater than the wild type (P < 0.005). The integrity of the cell fractions used for these studies was confirmed by assays for the periplasmic enzyme alkaline phosphatase. These assays showed no alkaline phosphatase activity in the supernatants from E. coli H10407 or MT13 at 4 and 8 h and comparable amounts of enzyme activity in supernatants at 24 h and in the periplasmic extracts from both strains at all time points.

Complementation of MT13 with a wild-type copy of gspD in trans on plasmid pMT44 restored LT secretion into culture supernatants to 74% of wild-type levels at 4 h, 76% at 8 h, and 89% at 24 h. These findings are similar to those reported when an epsD (gspD homologue) mutant of V. cholerae, strain N16961, was trans-complemented by a plasmid clone of wild-type epsD (34). Together, these data show that the type II secretion pathway is required for the secretion of LT in E. coli H10407.

The results of the bacterial viability/cell lysis assays showed no difference between E. coli H10407 and MT13. At 4 and 8 h, there was less than 1% cell lysis, and at 24 h there was 34% cell lysis, which could account for the presence of LT in the supernatant of MT13.

Biological Activity of LT.

To determine whether the LT produced by the mutant strain MT13 was biologically active, we used the Y1 adrenal cell bioassay. This assay is based on the observation that LT causes Y1 adrenal cells to change from their usual elongated spindle morphology to a more rounded shape. These morphological changes are indistinguishable from those induced by CT (18) and are concentration dependent. Y1 cells seeded in 96-well microtiter plates were incubated with cell-free supernatants and periplasmic extracts from E. coli H10407 and MT13. The morphological changes induced in the Y1 cells by these preparations were identical, indicating that the LT produced by strain MT13 was biologically active (Fig. 3).

A semiquantitative assay (27) using Y1 adrenal cells and 2-fold dilutions of the cell-free supernatants and periplasmic extracts was used to determine the amount of LT in these fractions. Using purified LT, we determined the sensitivity of the assay (amount of toxin that induced rounding in 50% of the cells) to be 49 pg/ml. This value was used to estimate the amount of LT in the test samples. The result of this assay confirmed those of the EIA, namely that the gspD mutant was capable of LT synthesis but was markedly defective in LT secretion (Figs. 2 and 3).