sequences encoding the enzymatically active A subunit of the cholera toxin. However, volunteer ... ing in vitro a ctxA deletion in the ctx operon cloned from .... The A- B+ strain thus ..... FIG. 4. Western blot of total protein from TnphoA insertion.
Vol. 60, No. 2
INFECTION AND IMMUNITY, Feb. 1992, p. 428-434
0019-9567/92/020428-07$02.00/0 Copyright © 1992, American Society for Microbiology
Cloning of a Gene (zot) Encoding a New Toxin Produced by Vibrio cholerae BERNADETTE BAUDRY,1'2 ALESSIO FASANO,lt JULIAN KETLEY,1t AND JAMES B. KAPERl.2* Center for Vaccine Development, Division of Geographic Medicine, Department of Medicine, University of Maryland School of Medicine,1 and Medical Biotechnology Center, Maryland Biotechnology Institute, University of Maryland,2 Baltimore, Maryland 21201 Received 22 July 1991/Accepted 8 November 1991 Live oral candidate cholera vaccines have previously been constructed by deletion of Vibrio cholerae sequences encoding the enzymatically active A subunit of the cholera toxin. However, volunteer studies have shown that these non-cholera toxin-producing strains still provoke mild to moderate diarrhea in some individuals. We recently reported the identification of a second toxin produced by V. chokrae which may be responsible for this residual diarrhea (A. Fasano, B. Baudry, D. W. Pumplin, S. S. Wasserman, B. D. Tall, J. M. Ketley, and J. B. Kaper, Proc. Natl. Acad. Sci. USA 88:5242-5246, 1991). This new toxigenic factor increases the permeability of rabbit ileal mucosa by affecting the structure of the intercellular tight junctions (zonula occludens). We now report the identification and cloning of the gene encoding this new toxin. This gene, named zot (for zonula occludens toxin), consists of a 1.3-kb open reading frame which could potentially encode a 44.8-kDa polypeptide. The location of the zot gene encoding the new toxin is immediately upstream of the ctx operon encoding cholera toxin.
Cholera, the severe diarrheal disease caused by Vibrio cholerae, still plagues many countries throughout the world. Although much effort and research have been invested in vaccine development, the construction of an attenuated live oral vaccine has proven unexpectedly difficult. The use of recombinant DNA techniques has led to the construction of several attenuated V. cholerae strains administered as live oral vaccines (16, 17, 23). These vaccine candidates were designed on the same basic principle, namely, the inactivation of the primary virulence factor of V. cholerae, cholera toxin. However, most of these strains were still capable of causing mild to moderate diarrhea when tested in volunteers. Cholera toxin is composed of one A subunit, which stimulates adenylate cyclase activity, and five identical B subunits, which are responsible for the binding of cholera toxin to the GM1 ganglioside receptor of intestinal mucosa (10, 11). The method of attenuation used in the construction of the previous vaccine candidates was the specific deletion of the major part of the gene encoding the A subunit (ctxA), resulting in an A- B+ strain (phenotypically CT-). V. cholerae CVD101 (A- B+) was obtained by first constructing in vitro a ctxA deletion in the ctx operon cloned from classic Ogawa strain 395. The ctx mutation was then reintroduced into the chromosome of 395 by allelic exchange (17). V. cholerae 395-Ni, also A- B+, was engineered in a similar way except that a ctxA deletion mutation was constructed in ctx sequences originally cloned from classic
of the volunteers (20). Under identical study conditions, strain 395-Ni caused significantly less diarrhea (in only 1 of 21 volunteers), but the production of other symptoms such as abdominal cramps, headache, and malaise, which were also seen with CVD101, was undiminished (14, 20). Our investigation into the basis for the different clinical responses observed with these two strains has resulted in the identification of a new toxic factor produced by V. cholerae. We recently reported that this second enterotoxin, Zot (zonula occludens toxin), increases the permeability of rabbit small-bowel mucosa by affecting the structure of the tight junction, or zonula occludens (7). In the present paper, we report the cloning and sequencing of the zot gene, which is located immediately upstream of the ctx genes on the chromosome. Expression of the zot gene product may explain the residual diarrhea seen in volunteers fed attenuated V. cholerae CVD101 and may possibly represent a new mechanism of bacterial diarrhea. MATERIALS AND METHODS Bacterial strains, plasmids, and media. V. cholerae strains used included classical Ogawa 395 and 395-Ni (23), classical Inaba 569B (received from J. Mekalanos, Harvard University) and CVD103HgR (19), El Tor Inaba JBK70 (16), El Tor Ogawa Texas Star (15), the nontoxigenic El Tor Ogawa 1196-78 (18), and V. cholerae non-O-1 strain 2076-79 (26). Cloning vectors included pBR322 (2), pUC19 (31), pJBK85 (17), M13mpl8 (24), and M13mpi9 (24). Plasmid pRK2013 (5) was used to mobilize recombinant plasmids into V. cholerae, and pRK751 was used to cure plasmids by incompatibility as previously described (17). pCVD18 (17) contains a 1.4-kb XbaI fragment encoding tetracycline resistance. Escherichia coli DHSa (BRL, Gaithersburg, Md.) or HB101 (3) was used for transformation and cloning, and DHSa F' was used for propagation of M13 derivatives. Bacteria were grown in L broth or on L agar. Antibiotics employed were ampicillin (200 ,ug/ml), tetracycline (30 ,ug/ml for E. coli, S
Inaba strain 569B and then recombined into the chromosome of Ogawa 395 (23). Both CVD101 and 395-Ni were markedly attenuated in volunteers compared with parent strain 395. However, despite the similarity of construction, the clinical responses observed with these strains were not identical. V. cholerae CVD101 caused mild to moderate diarrhea in 54% * Corresponding author. t Present address: Cattedra di Pediatria, Ospedale Pugliese, Universita di Catanzaro, Via Pio X88100, Catanzaro, Italy. t Present address: Department of Genetics, University of Leicester, Leicester LE1 7RH, United Kingdom.
428
V. CHOLERAE zot GENE
VOL. 60, 1992
pBB6
pBB68/168/268 pBB2681
P
c
I
i
P
C C
C
i
I
c
x I
I
C
P
C
I
.1
ctxA B
x x
C
C PC
C
l
l
4
Plasmid, clone, or mutant
E
x Tc
pJMK24 / pBB24
TABLE 1. Plasmids, M13 clones, and TnphoA mutants constructed in this study
E
c
E
-
X
zoxt FIG. 1. Restriction maps of the plasmid inserts used in the construction of CVD108 and the subcloning of the zot gene. Vectors for these plasmids are given in Table 1. Filled arrows represent the ctx genes, and the filled box represents the tet gene. The position and direction of transcription of the zot gene are marked by the open arrows. C, ClaI; E, EcoRI; P, PstI; X, XbaI.
pBB24 pBB241
pBB168 pBB268 pBB2681 pJMK23
,ug/ml for V. cholerae), kanamycin (50 ,ug/ml), polymyxin B (15 U/ml), and trimethoprim (50 ,ug/ml). Molecular genetic techniques. Standard recombinant DNA techniques as described by Maniatis et al. (22) were employed. The nucleotide sequence was determined by the dideoxy chain termination method using a Sequenase kit (USB, Cleveland, Ohio) and "S-dATP (Amersham, Arlington Heights, Ill.). The sequence was determined on the M13 clone JMK24 (Fig. 1), which is composed of the 2.7-kb XbaI-PstI fragment cloned in M13mpl9, and on JMK24R, which is composed of the same fragment in M13mpl8. Sequences were analyzed by using programs developed by the Genetics Computer Group at the University of Wisconsin (4). The programs employed included WORDSEARCH, MOTIFS, BESTFIT, and GAP. Insertion mutagenesis with TnphoA was performed as previously described, using SM10(pRT733) as the transposon donor (28). After mutagenesis, colonies harboring TnphoA fusions were pooled and the extracted plasmid DNA was transformed into HB101, therefore selecting for fusions located on the targeted recombinant plasmid. The insertion mutants were selected in the presence of ampicillin to ensure that the fusion was in the cloned insert and not in the exported ,-lactamase gene. Colonies carrying inserts of TnphoA were screened for phoA+ activity on L agar containing 20 ,ug of 5-bromo-4chloro-3-indolyl phosphate per ml. Construction of CVD108. A ctxA deletion mutant of V. cholerae 395 was constructed by first cloning the ctx operon on a 5.1-kb EcoRI-PstI fragment isolated from a chromosomal digest of V. cholerae 569B. The cloned ctx gene in pBR322 was designated pBB6 (Fig. 1). A 550-bp XbaI-ClaI fragment encoding the Al subunit of the cholera toxin was deleted to produce pBB68 by partially digesting pBB6 with ClaI to give linear molecules. The linear DNA fragments were isolated, treated with the Klenow fragment of DNA polymerase I, and ligated in the presence of XbaI linkers. Following digestion with XbaI, the fragments were religated and transformed into DH5a. The insert of pBB68 was subcloned into the IncP plasmid vector pJBK85 to give pBB268, and a selectable marker was added by cloning a tetracycline resistance gene (tet) from pCVD18 into the single XbaI site of pBB268 (Fig. 1). The resulting plasmid, pBB2681 (Fig. 1), was mobilized into V. cholerae 395 with
Characteristics
5.1-kb PstI-EcoRI fragment containing intact ctx and zot genes cloned into
pBR322, Tcr pBB68
l
zot
pBB241
Plasmids pBB6
X
C
429
pJMK25 M13 clones JMK24
JMK24R
Derivative of pBB6 in which 550-bp XbaIClaI fragment encoding ctxA has been deleted, CT- Zot+ Tcr 2.7-kb PstI-XbaI fragment containing zot and first 87 bp of ctxA cloned into pUC19, Zot+ Apr 1.5-kb ClaI-XbaI fragment from pBB24 cloned into pUC19, Zot+ Apr 4.7-kb PstI-EcoRI fragment of pBB68 cloned into pUC19, CT- Zot+ Apr 4.7-kb PstI-EcoRI fragment of pBB68 cloned into pJBK85, CT- Zot+ Cmr Derivative of pBB268 with tet gene inserted into XbaI site of ctxA, CTZot+ Cmr Tcr 2.7-kb PstI-XbaI fragment containing zot cloned into pCVD316, Zot+ Tcr 2.0-kb XbaI-EcoRI fragment from pBB68 cloned into pCVD316, Zot- Tcr 2.7-kb PstI-XbaI fragment of pJMK23 containing zot cloned into M13mpl9 2.7-kb PstI-Xbal fragment of pJMK23 containing zot cloned into M13mpl8
Mutants
C2-12, C2-32, C4-31 TnphoA mutants of pBB24, zot::TnphoA TnphoA mutants of pBB24, ctxA::TnphoA Cl-11, C4-22
pRK2013. Homologous recombination of the tet gene and flanking vibrio sequences into the chromosome of 395 was selected by introducing an incompatible plasmid, pR751 (Tpr), while maintaining selection for tetracycline and trimethoprim. Loss of both chromosomal copies of the Al subunit coding region was verified by DNA hybridization. Exchange of the ctx::tet copies with the /ctxA mutation carried by pBB268 was achieved by homologous recombination and selection of Tcs clones. The A- B+ strain thus obtained is CVD108. The plasmids, clones, and mutants constructed for this study are described in Table 1. Ussing chambers. Experiments were performed as previously described (12) with ileal segments of New Zealand White rabbits. Briefly, adult male rabbits (2 to 3 kg) were anesthetized by methoxyflurane inhalation and then sacrificed by air embolism. A 15-cm segment of ileum was removed, rinsed free of intestinal content, opened along the mesenteric border, and stripped of the muscular and serosal layers. Segments of this mucosa sheet were mounted in Ussing chambers (1.12-cm2 aperture) bathed in Ringer buffer (53 mM NaCl, 5 mM KCI, 30.5 mM Na2SO4, 30.5 mM mannitol, 1.69 mM Na2HPO4, 0.3 mM NaH2PO4, 1.25 mM CaCl2, 1.1 mM MgCl2, 25 mM NaHCO3). The bathing solution was maintained at 37°C and gassed with 95% 02-5% CO2. Potential difference, short-circuit current, and tissue ionic conductance (G,) were measured as previously described (8). G, values given in millimhos per square centimeter are equivalent to values given in millisiemens per
430
INFECT. IMMUN.
BAUDRY ET AL.
square centimeter (7). Culture supernatants were obtained by centrifugation at room temperature followed by filtration through a 0.45-pLm-pore-size membrane filter. At t = 0, 300 pJ of the filtrate was added to the mucosal side; 300 ,ul of the same sample was also added to the serosal side to preserve the osmotic balance. Each experiment used uninoculated L broth as a negative control. Western blots (immunoblots). Whole-cell extracts were prepared by centrifuging 1.5 ml of culture, suspending the pellet in denaturing sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis loading buffer (2 x buffer is 62.5 mM Tris HCl [pH 6.8], 10% glycerol [vol/vol], 5% ,3-mercaptoethanol [vol/vol], 2% SDS [wt/vol], and 0.01% bromophenol blue [wt/vol]), and boiling the samples for 5 min. Extracts were run on 12% SDS-polyacrylamide minigels and transferred onto nitrocellulose as previously described (29). The nitrocellulose was incubated for 30 min in blocking solution (phosphate-buffered saline, 1% Na-caseinate, 0.05% Tween 20, 0.05% NaH3) and then for 1 h with rabbit antibacterial alkaline phosphatase (diluted 1:1,000; a gift of David Low, University of Utah). The filters were washed twice for 30 min in PBS-0.05% Tween 20, incubated for 1 h with goat anti-rabbit immunoglobulin G conjugated with alkaline phosphatase (Sigma, St. Louis, Mo.), washed twice for 30 min each time in PBS-0.05% Tween 20, and exposed to the substrate {1 mg of Nitro Blue Tetrazolium, 0.5 mg of bromo-chloro-indolyl phosphate, and 20 I" of MgCl2 per 10 ml of Veronal buffer [1 vial of Barbital buffer (Sigma), 5.4 g of NaAc(3H20) in 400 ml, pH 9.6]}. Correlation of zot gene presence and Ussing chamber activity. A 575-bp StuI-AccI fragment internal to the zot gene was used as a hybridization probe to detect homologous sequences in other strains of V. cholerae. Culture supernatants of the strains probed for zot sequences were tested for Ussing chamber activity as described above. Nucleotide sequence accession number. The DNA sequence reported here has been submitted to GenBank (accession number M83563).
RESULTS Ussing chamber activity of CVD108. Attenuated V. cholerae strains CVD101 and 395-Ni are both derivatives of toxigenic strain 395 but do not give identical results when examined in Ussing chambers and volunteer studies. In Ussing chambers, CVD101 produced an immediate increase in G,, while 395-Ni produced no significant change in G, for nearly 100 min after addition of culture supernatants (7). CVD101 also produced significantly more diarrhea in volunteers than did 395-Ni (14, 20). Although both strains are derivatives of V. cholerae 395, they differ in the source of the cloned ctx genes used to construct the ctxA deletion; CVD101 was constructed by using cloned ctx genes from V. cholerae 395 (17), and 395-Ni was constructed by using ctx genes from strain 569B (23). To investigate the possibility that the difference observed in Ussing chambers and volunteers was due to the source of the cloned ctx genes (or sequences immediately flanking ctx), we constructed a third derivative of 395 designated CVD108. This strain was constructed as described above by using the same technique employed in the construction of CVD101 but utilized the same ctx DNA fragment from 569B that was used in the construction of 395-Ni. Culture supernatants of strains 395, CVD108, CVD101, and 395-Ni were tested for activity in Ussing chambers. Supernatants of 395, CVD108, and CVD101 (not shown)
time (hrs:mins)
0:00
0:.20
0:40
1:00
1:20
1:40
2:00
2:20
time (rs:mins) FIG. 2. Variation in G, with time as measured in the Ussing chambers. (A) Comparison of the two candidate vaccines, CVD108 and 395-Ni, with wild-type strain 395. (B) Complementation of 395-Ni activity by pBB2681. Open squares, medium control; open triangles, 395; closed circles, CVD108; closed triangles, 395-Ni; open circles 395-N1(pBB2681). Variation in G, is expressed as AG,, with AG, = G, at time x - G, at time zero. For clarity, the data are presented in two graphs, but the same experimental values for 395, 395-Ni, and the medium control are plotted in both panels.
provoked a rapid increase in G, starting within 20 min after addition of the samples (Fig. 2A). The culture supernatant of 395-Ni behaved like the negative (medium) control for an average of 90 min before conductance started to increase (Fig. 2A). As reported elsewhere (7), this increase in G, is due to a new toxin, Zot, which alters the structure of the tight junction. The results in Fig. 2A show that, unlike 395-Ni, CVD108 is unimpaired in its ability to produce this toxin, suggesting that the source of the ctx genes is not the cause of the difference observed in Ussing chamber activity. Localization of the gene encoding Zot. The difference in Ussing chamber activities seen with the culture supernatants of CVD108 and 395-Ni suggests that either a gene involved in the production of Zot is located close to the ctx genes and has been altered during the construction of strain 395-Ni or that a mutation independent of the ActxA genetic construction has altered the ability of 395-Ni to produce Zot. To examine the possibility that the gene encoding Zot is located near the ctx gene, we introduced plasmid pBB2681 (containing the ActxA::tet construct on a PstI-EcoRI fragment) into strain 395-Ni. When tested in Ussing chambers, the supernatant of V. cholerae 395Nl(pBB2681) gave an increase in G, similar to that observed with 395 or CVD108 supernatants (Fig. 2B). The ability of pBB2681 to restore Zot activity to 395-Ni suggests that a gene involved in the expression of Zot must be located in the region flanking the ctx genes. Plasmid pBB68 (containing the ActxA mutation on a 4.7-kb PstI-EcoRI fragment) was transformed into E. coli DH5a. The culture supernatant of the transformant was assayed in Ussing chambers and was found to induce an increase in tissue conductance (Table 2). This activity in E. coli suggests
V. CHOLERAE zot GENE
VOL. 60, 1992
TABLE 2. Activity of cloned zot genes in Ussing chambers Clone or strain" Localization of zot gene on region flanking ctx pBB68 pBB241 pBB24 V. cholerae 395 Medium
was subcloned in two halves by using the vector pCVD316
(9): pJMK23, containing the 2.7-kb XbaI-PstI fragment
AG, (mS/cm2)" 6.5 7.8 8.5 6.9 1.5
± ± ± ± ±
1.4 (2) 2.3 (3) 0.6 (4) 0.9 (4) 0.4 (4)
Evaluation of upstream and downstream ctx regions for zot gene localization pJMK23 pJMK25 V. cholerae 395 Medium
6.9 2.0 8.0 2.5
± ± ± ±
1.5 (3) 1.0 (3) 1.0 (3) 0.5 (3)
Effect of TnphoA mutants on zot expression C1-i C4-22 C2-12 C2-32 C4-31 V. cholerae 395 Medium
8.8 7.3 3.5 1.1 3.4 7.0 2.2
± ± ± ± ± ± ±
2.0 (2) 1.7 (2) 1.0 (3) 0.5 (2) 0.7 (2) 0.6 (4) 0.3 (4)
upstream of the ctx gene, and pJMK25, containing the 2.0-kb
XbaI-EcoRI downstream fragment. Only pJMK23 retained the ability to restore the production of Zot in 395-Ni (Table 2). Therefore, the gene involved in the production of Zot is localized in the 2.7-kb region upstream of the ctx operon. Sequence of the region upstream of ctx. The DNA sequence of the 2.7-kb XbaI-PstI insert contained in pJMK23 revealed a 1.3-kb open reading frame (ORF) which is shown in Fig. 3. The gene encoding this ORF, designated zot, is located immediately upstream of the ctx gene and is transcribed in the same direction. Only 127 bp separate the zot stop codon from the ctxA start codon (Fig. 3). The predicted primary translation product from the first methionine of zot would be a 44.8-kDa polypeptide containing 399 amino acid residues with a predicted pI of 8.5. Although a classic signal peptide sequence is lacking, a potential signal peptidase cleavage site (30) is found after the first 18 amino acids. The 3' end of a second large ORF, with a 5' end beyond the PstI site, was found immediately upstream of the zot gene (data not
shown). Mutagenesis of the zot gene. To confirm that the 1.3-kb ORF encoded Zot, TnphoA mutagenesis was performed on HB101(pBB24) (the sequenced 2.7-kb insert of pJMK23 cloned in vector pUC19). Four independent mutagenesis experiments yielded eight phoA+ clones carrying an insertion in the cloned Vibrio DNA. These clones were screened by restriction enzyme profiles, and only two classes of restriction patterns were observed. Five independent clones were chosen, and culture supernatants from each clone were tested in Ussing chambers. The two mutants of one class
a Clones pBB68, pBB241, pBB24, pJMK23, and PJMK25 were tested in an E. coli DH5a background; TnphoA mutants were tested in E. coli HB101. V. cholerae 395 served as a positive control. I Values are means ± standard error, with number of repetitions in parentheses.
that the 4.7-kb insert of pBB68 contains the structural gene for the Zot factor rather than a regulatory gene involved in expression of Zot, although we cannot absolutely rule out the latter possibility at this time. The 4.7-kb insert of pBB68 @00000
*000
000000
GGCTATCGATATGCTGTCTCCTCAATACAGCGCTTTCTGGCTATCGTGCTTCAGGCTTTGATGACCCGTTTCGCCCTGCGAGCGTIAAACCT 121 241
431
M S I F I H H G A P G S Y K T S c; A4 L v L R L L P A I K S G R H I I T N V R G L ATGAGTATCTTTATTCATCACGGCGCGCCAGGCTCTTATAAAACGTCCGGGCATTATGGCTTCGTCTGCTGCCGGCGATTAAGTCAGGCCGTCACATCATCACGAATGTGCGAGC'CTTA HN A R N L E R I A K Y L K M D V S D I S I E F I D T D H P D G R L T M A R FN AACCTTGMAAATAGCTAAGTACTTAAAAATGGACGTCTCAGACATCAGTATCGAGTTTATTGATACAGACCATCCAGACGGTCGCTTAACGATGGCGCGTTTTTGGCACTGGGCGAGA K D A F L F I D E C G R I V P P R L T AT N L K A L D T P P D L V A E D R P E S
120 240 360
AAGGACGCGTTTCTCTTTATTGATGAATGTGGTCGCATCTGGCCGCGGCTGACGACAATTTAAAGGCGC TCGACACGCCGCCGGATTTGGTCGCAGAGGATAGGCCTGAGAGC
480
481
E V A F D M H R H H G W D I C L T T P N I A K V H N M I R E A A E I G Y R H F TTTGAGGTGGCTTTTGACATGCATCGTCACCACGGCTGGGATATCTGCCTAACCACGCCTAACATTGCCAAAGTGCACAACATGATAAGAGAGGCGGCGGAGATAGGGTATCGCCACTTT
600
601
AACCGCGCCGGTGGGGCTAGGGCAATTACCCTGACCACCCACGATGCAGCCAACTCTGGACAGATGGATTCGCACGCGCTGACACGCCAAGTCAAAAAAMTTCCAAGTCCGATT
361
F
N
F
721
T
K M Y
V
A
G S
L T
G
T
A K F T
G
T
K A
L R
T D
T
H
D
T N A
A A G
T
N A
S
G
Q M D S H A L T R Q V K K I P S P
L W K
D
R
K
I
L
F
L
F
G
M V
F
L M
F
S
I
Y
F
Y
G
L
H
D
N
P
I
F
T
G
G
N
D
A
T
I
E
S
E
Q S
E
P
Q
S
K
A
T
A
G
N
A
V
G
S
K
960
Q D G F V T V G D E R Y R L V D N L D I P Y R G L GCTGCTCCTGCGTCTTTTGGTTTTTGTATTGGTCGGCTTTGTGTCCAAGATGGTTTTGTCACTGTTGGTGATGAGCGTTATCGCCTCGTAGACAATTTGGACATTCCTTATCGTGGTCTA
1080
A
P
A
S
F
G
F
C
I
G
R
L
C
V
C4-, TTTTTGATTTTTTGATTTTTGATT-Kx'xGATTTTTAsTTT M V
1441
840
A
N A T G H H I Y K D T L T V F F E T E S G S V P T E L F A S S Y R Y KV L P L P 1081 TGGGCGACAGGTCATCACATTTACAGACCTCGTGTTT TTGAAACCG AGAGTG GCAGCGTCCCAACAGAGC.TGTTTGCATCGAGCTACCGCTACAA(GCTACCGTTACCG C2-12 D F N H F V V F D T F A A Q A L W V E V K R G L P I K T E N D K K G L N S IF * 1201 G VTTTCAATCACTTTGTGGII GFATACCTTTGCAGCAAGCGTGGGTAGAAGAA TTTACCGATAAAAACAGAA T AA GACTTAGTATATTTTA 1321
720
TCGTTTTACGGCTTACACGACAATCCAATTTTACAGGGGGAAATGATGCAACTATCGAGTCAAGATCACCTCAGTCAAAGGCTACTGCTGGGAATGCTGTCGGGAGCAAGGCG A
961
A
TTTAAGATGTACGCAGACACAAGC ACAGGACACGATGGCCGGAACGGCGCTGTGAAAGACAGAACTTTTCTTGTTCGGCATGGTTTTTTTGATGTTCTCTTAT S
841
R
K
I
I
C2-3 TAmGnCAsCUTTAmATTGTTTGATCATTTATTTTCTGTTACAAGGA
1200 1320
1440
F
CATTATATGGTAAAGATAATATTT 1464 ,ctxA
FIG. 3. Nucleotide sequence of zot showing position of the TnphoA insertions (C2-12, C4-31, and C2-32) and the start of the ctxA gene. Open dots indicate possible -35 and -10 promoter sequences, and black dots indicate a potential ribosome-binding site. A possible cleavage site for signal peptidase is indicated by the vertical arrow.
432
INFECT. IMMUN.
BAUDRY ET AL.
1
2
3 4 5 6
20092.5_ 69-
4 6-
kDa
3021.514.3 -
FIG. 4. Western blot of total protein from TnphoA insertion
TABLE 3. Correlation of zot gene presence and Ussing chamber activity Strain
AG, (mS/cm2)a
zot geneb
V. cholerae 395 JBK70 Texas Star CVD103HgR 1196-78 2076-79 E. coli HB101
6.9 8.5 7.4 7.7 2.0 1.7 2.4
± 1.7 (9) ± 2.5 (3) ± 2.0 (3) ± 1.5 (3)
+ + + +
± 0.2 (3) ± 0.8 (3) ± 1.0 (9)
a Values are means ± standard error, with number of replicates given in parentheses. b +, hybridization of zot gene probe to indicated strain; -, no hybridization.
from Brazil which naturally lacks ctx sequences and is avirulent in volunteers (18). The other strain, V. cholerae non-O1 2076-79, also lacks ctx but produces a heat-stable enterotoxin and causes diarrhea in volunteers (26).
mutants. Lanes: 1, rainbow molecular weight markers (Amersham);
2, HB101(pBB24); 3, HB101 (Cl-11); 4, HB101 (C2-12); 5, HB101 (C2-32); 6, HB101 (C4-22).
(Cl-11 and C4-22) had retained Zot activity, while the three clones of the other class (C2-12, C2-32, and C4-31) had lost the ability to increase G, (Table 2). The precise locations of the TnphoA insertions were determined by sequencing the junction between the Vibrio DNA and the TnphoA. In the two mutants that retained Zot activity, TnphoA was found to be inserted in the truncated portion of the ctxA gene remaining in pBB24. In mutants C2-12, C2-32, and C4-31, the insertions were located in the 1.3-kb ORF just upstream of the ctxA gene, all within 150 bp from the end of the ORF (Fig. 3). On the basis of these data, a smaller, 1.5-kb ClaI-XbaI fragment was cloned into pUC19 to yield plasmid pBB241 (Fig. 1). The culture supernatant of DH5a(pBB241) was tested in Ussing chambers and found to increase the G, (Table 2). All of these findings indicate that the 1.3-kb ORF located just upstream of the ctxA gene encodes Zot. Four TnphoA insertion mutants were analyzed for PhoA fusion proteins by immunoblotting with rabbit antibacterial alkaline phosphatase (Fig. 4). Two of these mutants, C2-12 and C2-32, carry an insertion in the 1.3-kb ORF, while the other two, Cl-11 and C4-22, contain an insertion in the residual part of the ctxA gene. In the case of Cl-11 and C4-22, the apparent molecular weight of the fusion product was ca. 48 kDa. This size, just slightly larger than that of bacterial alkaline phosphatase, is consistent with the position of the TnphoA inserts in the 5' end of the ctxA structural gene. C2-12 and C2-32 both expressed a fusion product of ca. 92 kDa, which is similar to the size of the fusion protein predicted from the zot sequence data and the position of the TnphoA insertions within zot. Correlation of zot gene presence and Ussing chamber activity. Table 3 shows the correlation of Ussing chamber activity and the presence of zot gene sequences. Four of the six strains produced AG, values of >6 mS/cm2, and all four strains contained sequences homologous to zot. The two strains which did not contain zot sequences had essentially background levels of Ussing chamber activity. One of these strains, V. cholerae 01 strain 1196-78, is a sewage isolate
DISCUSSION We recently reported the existence of a second potential enterotoxin in V. cholerae (7). This toxin, named Zot for zonula occludens toxin, affects the structure of epithelial tight junctions. Induction of diarrhea in volunteers by attenuated V. cholerae CVD101 but not 395-Ni correlated with an immediate increase in G, in the Ussing chamber assay with CVD101 but not with 395-Ni. Because of the difference in activity of CVD101 and 395-Ni, we initially constructed CVD108, which was derived from V. cholerae 395 by using cloned ctx sequences from strain 569B. The increase in G, induced by CVD108 was equivalent to that induced by CVD101, and when fed to adult volunteers, CVD108 produced diarrhea in 9 of 13 subjects (27). In the present study, the gene encoding Zot was identified and localized to a site immediately upstream of the ctx genes on the chromosome. For the classic Inaba strain 569B, from which this clone was derived, the termination codon of the zot gene overlapped the initial TTTTGAT repeat of the ctx promoter. This repeated sequence has been shown by Miller et al. to be important in the expression of ctx by serving as a binding site for the ToxR transcriptional activator protein (25). Our preliminary data suggest that expression of the zot gene does not require ToxR, but further work must be done to deal with this possibility definitively. The predicted amino acid sequence of Zot showed no significant homology to sequences stored in the GenBank data base (Release 67), and analysis of the sequence with the program MOTIFS showed no significant homology to protein patterns found in the PROSITE Dictionary of Protein Sites and Patterns. The size of the Zot protein predicted from the DNA sequence is 44.8 kDa. The size of the Zot-PhoA fusion protein and the placement of potential promoter and ribosomal-binding sites immediately upstream of the first ATG in the 1.3-kb ORF suggests that this entire ORF is transcribed and translated. In our earlier study (7), differential filtration of the culture supernatant showed that the active protein was in the 10- to 30-kDa fraction. This discrepancy may indicate that the 44.8-kDa protein is processed to a final size of less than 30 kDa, or it may be due to imprecision in the sizing of the protein in the ultrafiltration step. At least one posttranslational processing step is suggested, as Ussing chamber activity was readily detected in V. cholerae culture super-
V. CHOLERAE zot GENE
VOL. 60, 1992
natants and a potential signal peptidase cleavage site was noted after the first 18 amino acid residues. Interestingly, Ussing chamber activity was also detected in culture supernatants of E. coli containing the cloned zot gene, but at present, we cannot rule out the possibility that the extracellular activity in E. coli is due to autolysis. Our ongoing studies concerning purification of the Zot protein and expression of the zot gene will yield more definitive data on these issues. The unusual mode of action of Zot makes it an inherently interesting polypeptide. It is, to our knowledge, the first bacterial factor to be described which is capable of reversibly altering tight junctions. The ability to loosen tight junctions has been reported for Clostridium difficile toxin A (13) and for influenza and vesicular stomatis viruses (21), but these activities are not reversible, nor have they been correlated with diarrheagenicity for these organisms. The sequence of the C. difficile toxin A (6) has recently been determined, but no similarity with our zot sequence was found. We have examined other diarrheagenic organisms such as Shigella, Salmonella and Yersinia species and diarrheagenic E. coli under low- and high-stringency hybridization conditions but have found no sequences homologous to zot in these species (1). The identification of the zot gene is a potentially highly significant step in the development of a safe live oral vaccine against cholera. This toxin may be responsible, at least in part, for the residual diarrhea seen in volunteers fed attenuated V. cholerae CVD101 and CVD108. V. cholerae CVD101 and CVD108 produce sufficient Zot to elicit an immediate response in the Ussing chambers, while 395-Ni, which causes little or no diarrhea in volunteers, produces either a less-active toxin or smaller amounts of Zot (7). It is possible that low levels of Zot are responsible for the other undesirable symptoms such as abdominal cramps, malaise, vomiting, and headaches seen with all three strains (CVD101, CVD108, and 395-Ni). If the potential to cause diarrhea is confirmed in subsequent studies, then Zot represents a completely new mechanism of infectious diarrhea. We are currently in the process of constructing isogenic strains deleted for both the ctxA and zot genes. These constructions should answer many questions concerning Zot and the ability of this toxin to cause diarrhea and could potentially yield a safe and effective live oral cholera vaccine. REFERENCES 1. Baudry, B., K. Gicquelais, and J. B. Kaper. Unpublished data. 2. Bolivar, F., R. L. Rodriguez, P. J. Greene, M. C. Betlach, H. L. Heynecker, H. W. Boyer, J. H. Crosa, and S. Falkow. 1977. Construction and characterisation of new cloning vehicles. Gene 2:95-113. 3. Boyer, H., and D. A. Roulland-Dussoix. 1969. A complementation analysis of the restriction and modification of the DNA in Escherichia coli. J. Mol. Biol. 41:459-472. 4. Devereux, J., P. Haeberli, and 0. Smithies. 1984. A comprehensive set of sequence analysis programs for the VAX. Nucleic Acids Res. 12:387-395. 5. Ditta, G., S. Stanfield, D. Corbin, and D. R. Helinski. 1980. Broad host range DNA cloning system for Gram-negative bacteria: construction of a gene bank of Rhizobium meliloti. Proc. Natl. Acad. Sci. USA 77:7347-7351. 6. Dove, C. H., S.-Z. Wang, S. B. Price, C. J. Phelps, D. M. Lyerly, T. D. Wilkins, and J. L. Johnson. 1990. Molecular characterization of the Clostridium difficile toxin A gene. Infect. Immun. 58:480-488. 7. Fasano, A., B. Baudry, D. W. Pumplin, S. S. Wasserman, B. D. Tall, J. M. Ketley, and J. B. Kaper. 1991. Vibrio cholerae
8. 9. 10. 11. 12.
13.
14.
15.
16.
17. 18.
19.
20.
21.
22.
23.
24. 25.
26.
27. 28.
433
produces a second enterotoxin, which affects intestinal tight junctions. Proc. Natl. Acad. Sci. USA 88:5242-5246. Field, M., D. Fromm, and I. McColl. 1971. Ion transport in the rabbit ileal mucosa. Part I. Na and Cl fluxes and short circuit current. Am. J. Physiol. 220:1388-1396. Galen, J. E., and J. B. Kaper. Unpublished data. Gill, D. M. 1976. The arrangements of subunits in cholera toxin. Biochemistry 15:1242-1248. Gill, D. M., and R. Meren. 1978. ADP-ribosylation of membrane protein catalysed by cholera toxin: basis of the activation of adenylate cyclase. Proc. Natl. Acad. Sci. USA 75:3050-3054. Guandalini, S., A. Fasano, M. Migliavacca, M. C. Verga, P. Mastrantonio Gianfrilli, A. Ferrara, M. Alessio, B. Malamisura, P. Galati, A. Pantosti, and A. Rubino. 1988. Pathogenesis of postantibiotic diarrhoea caused by Clostridium difficile: an in vitro study in the rabbit intestine. Gut 29:598-602. Hecht, G., C. Pothoulakis, J. T. LaMont, and J. L. Madara. 1988. Clostridium difficile toxin A perturbs cytoskeletal structure and tight junction permeability of the cultured human intestinal epithelium monolayers. J. Clin. Invest. 82:1516-1524. Herrington, D. A., R. H. Hall, G. Losonsky, J. J. Mekalanos, R. K. Taylor, and M. M. Levine. 1988. Toxin, toxin-coregulated pili, and the toxR regulon are essential for Vibrio cholerae pathogenesis in humans. J. Exp. Med. 168:1487-1492. Honda, T., and R. A. Finkelstein. 1979. Selection and characteristics of a Vibrio cholerae mutant lacking the A (ADPribosylating) portion of the cholera enterotoxin. Proc. Natl. Acad. Sci. USA 76:2052-2056. Kaper, J. B., H. Lockman, M. M. Baldini, and M. M. Levine. 1984. Recombinant nontoxigenic Vibrio cholerae strains as attenuated cholera vaccines candidates. Nature (London) 308: 655-658. Kaper, J. B., H. Lockman, M. M. Baldini, and M. M. Levine. 1984. A recombinant live oral cholera vaccine. Bio/Technology 2:345-349. Levine, M. M., R. E. Black, M. L. Clements, L. Cisneros, A. Saah, D. R. Nalin, D. M. Gill, J. P. Craig, C. R. Young, and P. Ristaino. 1982. The pathogenicity of nonenterotoxigenic Vibrio cholerae serogroup 01 biotype El Tor isolated from sewage water in Brazil. J. Infect. Dis. 145:296-299. Levine, M. M., J. B. Kaper, D. Herrington, J. Ketley, G. Losonsky, C. 0. Tacket, B. Tall, and S. Cryz. 1988. Safety, immunogenicity, and efficacy of recombinant live oral cholera vaccines, CVD 103 and CVD 103-HgR. Lancet ii:467-470. Levine, M. M., J. B. Kaper, D. Herrington, G. Losonsky, J. G. Morris, M. L. Clements, R. E. Black, B. Tall, and R. Hall. 1988. Volunteer studies of deletion mutant of Vibrio cholerae 01 prepared by recombinant techniques. Infect. Immun. 56:161167. Lopez-Vancell, R., G. Beaty, E. Stefani, E. E. Rodriguez-Boulan, and M. Cereijido. 1984. Changes in the paracellular and cellular ionic permeabilities of the monolayers of MDCK cells infected with influenza or vesicular stomatitis viruses. J. Membrane Biol. 81:171-180. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular cloning: a laboratory manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. Mekalanos, J. J., D. J. Swartz, G. D. N. Pearson, N. Harford, F. Groyne, and M. de Wilde. 1983. Cholera toxin genes: nucleotide sequence, deletion analysis and vaccine development. Nature (London) 306:551-557. Messing, J. 1983. New M13 vectors for cloning. Methods Enzymol. 101:20-78. Miller, V. L., R. K. Taylor, and J. J. Mekalanos. 1987. Cholera toxin transcriptional activator ToxR is a transmembrane DNA binding protein. Cell 48:271-279. Morris, J. G., Jr., T. Takeda, B. D. Tall, G. A. Losonsky, S. K. Bhattacharya, B. D. Forrest, B. A. Kay, and M. Nishibuchi. Experimental non-0 group 1 Vibrio cholerae gastroenteritis in humans. J. Clin. Invest. 85:697-705. Tacket, C. 0. (University of Maryland). Personal communication. Taylor, R. K., C. Manoil, and J. J. Mekalanos. 1989. Broad-
434
BAUDRY ET AL.
host-range vectors for the delivery of TnphoA: use in genetic analysis of secreted virulence determinants of Vibrio cholerae. J. Bacteriol. 171:1870-1878. 29. Towbin, H., T. Staehelin, and J. Gordon. 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci.
INFECT. IMMUN.
USA 76:4350-4354. 30. Von Heijne, G. 1986. A new method for predicting signal sequence cleavage site. Nucleic Acids Res. 14:4683-4690. 31. Yanisch-Perron, C., J. Vieira, and J. Messing. 1985. Improved M13 phage vectors and host strains: nucleotide sequences of the M13mpl8 and pUC19 vectors. Gene 33:103-119.
