Biological and immunological characterization of a cloned cholera toxin-like enterotoxin from Salmonella typhimurium

Microbial Pathogenesis 1990 ; 9 : 315-329

Biological and immunological characterization of a cloned cholera toxin-like enterotoxin from Salmonella

typhimurium Rajendra Prasad,' Ashok K . Chopra,' Johnny W . Peterson,' Roser Pericas 2 and Clifford W . Houston' 'Department of Microbiology, University of Texas Medical Branch, Galveston, Texas 775502782, U.S .A . and 2 Hospital San Pau, Microbiologie, Barcelona 08025, Spain

(Received April 9, 1990 ; accepted in revised form August 7, 1990)

Prasad, R . (Dept of Microbiology, University of Texas Medical Branch, Galveston, Texas 775502782, U .S .A .), A . K . Chopra, J . W. Peterson, R . Pericas and C . W . Houston . Biological and immunological characterization of a cloned cholera toxin-like enterotoxin from Salmonella typhimurium . Microbial Pathogenesis 1990; 9 : 315-329 . A chromosomal DNA fragment, encoding an enterotoxin gene of Salmonella typhimurium Q1, was cloned into bacteriophage EMBL3 and plasmid vector pBR322 . The recombinant clones , B8 and pC1 were identified using a synthetic oligonucleotide probe made to the B subunit region of the cholera toxin gene (ctx) . Cell lysates of Escherichia coli VCS257[a.B8] induced fluid secretion in rabbit intestinal loops, while lysates of E. coli DH5a[pC1 ] failed to elicit an enterotoxic response in this model . Both lysates and partially purified preparations elongated Chinese hamster ovary (CHO) cells, elevated cellular cAMP and PGE 2 , and bound to ganglioside G M,, . The biological activity associated with the cloned enterotoxin was neutralized by monospecific antiserum to cholera toxin (CT) . Immunoblots of pC1 and ,B8 lysates probed with anti-CT, exhibited a 30 kDa protein similar to that of pJM17, which carried the ctx gene . Under non-dissociating conditions, anti-CT immunoblots of the same lysates revealed two proteins, one corresponding in size to the holotoxin and the other to CT-A . When analysed by DNA-directed protein synthesis in vitro, both pC1 and %B8 DNA expressed two unique proteins (30 and 11 kDa) similar to that of pJM1 7 . Key words : Salmonella enterotoxin ; G M11 ganglioside ; ELISA; cholera toxin ; cAMP ; PGE2 .

Introduction Salmonella species cause a wide spectrum of human diseases ranging from self-

limited watery diarrhea to inflammatory diarrhea, bacteremia, enteric fever, and focal extraintestinal infections . Clearly, Salmonella typhimurium is a sophisticated pathogen that is well adapted to survive and grow in vivo . Despite extensive genetic studies on Salmonella, very little is known about the genetic determinants responsible for the virulence of this organism . Recently, there has been increasing interest in the study of the mechanisms involved in the pathogenesis of Salmonella infections . Several investigators 1-5 correlated the presence of a large plasmid in various Salmonella species to the enhanced virulence . Recently, a genetic determinant was identified in S . typhimurium SR-11 that conferred the ability of a non-invasive strain of S. typhimurium DB4673 6 to enter cultured mammalian cells . Similarly, Fields et al. 7 isolated a gene 0882-4010/90/110315+15 $03 .00/0

© 1990 Academic Press Limited



31 6

R . Prasad et al .

from S. typhimurium that was required for its virulence and intracellular survival in the host . Enterotoxins are one of several virulence factors produced by microbial pathogens .' In Vibrio cho/erae and Escherichia coil, it has been clearly demonstrated that enterotoxins play an important role in diarrheal disease . 9' 10 Though the role of an enterotoxin in salmonellosis is still controversial, substantial evidence has now accumulated to support the concept that Salmonella strains elaborate an enterotoxin ." -- " A heat-labile Salmonella enterotoxin has been purified partially and characterized ." , " The biological activities associated with the Salmonella enterotoxin, i .e ., fluid secretion in ligated rabbit intestinal loops and vascular permeability factor activity in rabbit skin in vivo, as well as elongation of Chinese hamster ovary (CHO) cells in vitro, were shown to be neutralized by monospecific antiserum made to CT ." The expression of CT-related antigen in several strains of Salmonella has also been reported using .20,21 However, these fluorescent antibodies to CT as well as by immunoblots . 17 investigators 20 failed to demonstrate CT-related antigen by ELISA with either G N11 ganglioside or anti-CT as an anchor . Interestingly, immunoblot analysis revealed .20 greater amounts of CT-related antigen in strains that were considered avirulent Extensive characterization of V. cholerae enterotoxin and E. coli enterotoxin LT1 revealed that the two toxins were immunologically cross-reactive and had a similar subunit structure ." Finkelstein et al." observed that Salmonella enterotoxin was similar to both cholera toxin and the E. coil enterotoxin, consisting of two subunits (A and B) . Our laboratory has demonstrated some DNA sequence homology between the chromosomal DNA of Salmonella strains and the cholera toxin gene .24, 25 Recently, Fernandez et al. 26 reported the molecular cloning of an LT-like enterotoxin gene of S . typhi into E. co/i. In this paper, we attempted to characterize partially a CT-like enterotoxin from S . typhimurium Q1 after cloning it into E. coil.

Results Screening of the S . typhimurium gene library We used a 32 P-labeled synthetic oligonucleotide (21 -mer) made to the B subunit of the ctx gene (position 1625-1645 of the nucleotide sequence) as a probe to screen the genomic library of S. typhimurium Q1 constructed in bacteriophage EMBL3 . Positive plaques were localized, picked, and stored, as described by Maniatis et al." We also cloned an EcoRl-digested chromosomal DNA fragment of S. typhimurium Q1 into pBR322, because it reacted strongly with the above mentioned synthetic oligonucleotide probe in Southern blots . Finally, two clones, namely pC1 and %B8, were selected for characterization of the gene product(s) . The sizes of the Salmonella chromosomal DNA fragments in pC1 and AB8 were approximately 5 .0 kb and 14 .5 kb, respectively . Biological nature of the cloned Salmonella enterotoxin The biological activity of the cloned Salmonella enterotoxin in E. coil was examined in rabbit intestinal loops and on CHO cells . Table 1 indicated that the cell lysates of E. coil VCS257, infected with the recombinant phage DNA [ . B8], resulted in the secretion of 1 .5 ml of fluid per cm of intestinal loop . Cell lysates from E. co/i DH5a[pC1], on the other hand, failed to elicit an enterotoxic response in the rabbit intestinal loop test . However, in a CHO cell elongation assay, cell lysates of both pC1 and 268 resulted in elongation of CHO cells [Fig . 1 (C) and (D)] . Lysates prepared from E. coil transformed with the plasmid [pC1 ] exhibited comparatively lower CHO

CT-like Salmonella enterotoxin

Table 1 otoxin

317

Biological and immunological properties of the cloned Salmonella enter-

Whole cell lysates co/i VCS257[EMBL3] coli VCS257[,n.B8] E. coli DH5a[pBR322] E. coli DH5a[pCl] E. coli VCS257[ : B8]+preimmune serum E. co/i VCS257[tiB8]+immune serum (CT) Cholera toxin (100 ng/ml)

Fluid accumulation' (ml/cm)

E.

0

E.

1 .5±0 .5`

CHO cell elongation'

G,,-ELISA (CT equiv . ng/ml )

0

0

64

70

0

0

0

0

16 64

14 nda nda 100

1 .8±0 .4° 0

2 .0±0 .5`

0

128

Fluid accumulation in rabbit intestinal ligated loops . ° Numbers indicate the highest dilution of the toxin causing more than 50% of the CHO cells to elongate . Mean ± SD from eight animals . ° Mean +SD from two animals . and = not determined . a

cell elongation titers, as compared to that of A.B8 phage lysates (Table 1) . Furthermore, unlike pJM1 7 (carrying the ctx gene), 28 treatment of cell lysates of pC1 and 2B8 with trypsin did not increase the biological activity of the cloned Salmonella enterotoxin (data not shown) . The cloned Salmonella enterotoxin from lysates of E. coli DH5a[pC1 ] was partially purified by P-11 phosphocellulose chromatography (Fig . 2) . The fractions (87-92) eluting from the column with 0 .2 M sodium phosphate buffer (pH 7 .4) exhibited positive G M ,-ELISA reactions using antibodies to CT . These ELISA-positive fractions after dialysis also elongated CHO cells [Fig . 1 (E)] . The column fractions (88 and 89) which exhibited the highest G M1 -ELISA values revealed the presence of a CT crossreactive protein band at 30 kDa in an immunoblot analysis (inset Fig . 2) . Further evidence that these clones [pC1 and A138] encoded a CT-like enterotoxin was obtained from the neutralization studies . When the cell lysates or partially purified preparations of these clones were mixed with immune serum directed against pure CT, we observed complete neutralization of Salmonella enterotoxin-induced fluid secretion and of CHO cell elongation [Table 1 and Fig . 3 (B)] . In addition, the CHO cell elongation activity was neutralized by anti-CTB and G M1 -ganglioside but not by anti-CTA (data not shown) . Enterotoxin preparations mixed with preimmune serum, on the other hand, exhibited normal elongation of the CHO cells [Fig . 3(A)] and accumulated fluid (1 .8 ml/cm) in rabbit intestinal loops (Table 1) . To study the mechanism of action of the Salmonella enterotoxin further, we measured the cyclic adenosine monophosphate (cAMP) and prostaglandin (PGE 2 ) levels of the CHO cells treated with the Salmonella enterotoxin . Our data indicated that the crude and partially purified preparations of the cloned Salmonella enterotoxin, like CT, 29,30 increased the intracellular cAMP concentration and the release of PGE 2 from the CHO cells (Fig . 4) . Structure and antigenic nature of the Salmonella enterotoxin The cloned Salmonella enterotoxin was examined by immunoblot analysis and by in vitro DNA-directed transcription/translation of isolated plasmid and phage DNA encoding the toxin gene . Immunoblot analysis . Whole cell lysates of E. co/i DH5a[pC1 ] and E. coil VCS257[),B8] were analysed by SDS-PAGE . The separated proteins were transferred

318

R . Prasad et al .

Fig . 1 . Effect of the cloned Salmonella enterotoxin on morphology of cultured CHO cells . The tissue culture cells were treated with the cell lysates of (A) E. co/i DH5a[pBR322] ; (B) E. co/i VCS257[EMBL3] ; (C) E. coli DH5x[pC1] ; (D) E. co/i VCS257[%B8] ; (E) biologically active pool of pC1 prepared by P-11 phosphocellulose chromatography ; (F) cholera toxin . The CHO cells were fixed and stained with Giemsa"s stain . Bar represents 50 tim .

onto a nitrocellulose membrane, which then was probed with affinity-purified antibodies to CT . Clearly, a 30 kDa polypeptide in the pC1 and ).B8 lysates reacted with anti-CT (Fig . 5, lanes 6, 7 and 8) . A 30 kDa protein band, representing the A subunit of CT, was detected also in the whole cell lysates of pJM17 carrying the complete ctx gene (Fig . 5, lane 9) . However, antibodies specific to CTA, but not to the CTB,



CT-like Salmonella enterotoxin 3 .5

319 2 .0

/ 84 85 88 89

3 .0

a 2 .5 ~

a

I

I

I

1 .0

v

1 .5

I

0

1

L

a

I

s

J 0

10

60

70

J W

II

0.5

I I'

N C1

0 .5

O 0

0 a LO

I

L

1.0



I I

w

2 .0

m

0 °

1 .5

=

I \ ~ I \

80 90 Fraction number

100

110

120

0.0 130

Fig . 2 . Elution profile of the cell lysates of E. coli DH5a[pC1 ] on a P-11 phosphocel I u lose column . The cell lysate dialysed against 10 mm sodium phosphate buffer (pH 7 .0) was applied to the column, which was equilibrated with the same buffer . The column was washed with 10 mm sodium phosphate buffer (pH 7 .0) until the absorbance of the eluted fractions reached the baseline value . The bound proteins then were eluted with 0 .2 M sodium phosphate buffer (pH 7 .4) as indicated by the arrow . Fractions (4 ml) were collected and the presence of Salmonella enterotoxin antigen was examined in the fractions by a G N„-ELISA using affinity-purified antibodies to CT . The inset shows an immunoblot of G ti„ ELISA-positive fractions . The arrowhead indicates a 30 kDa CT-cross reactive protein band .

detected a 30 kDa protein band in whole lysates of pC1, ,!B8, and pJM17 (data not shown) . No comparable protein band was detected in the control lysates of E. co/i DH5a[pBR322] (Fig . 5, lanes 1 and 2) or E. co/i VCS257 infected with 2EMBL3 (Fig . 5, lane 3) . There was no evidence of an 11 kDa protein band (corresponding to the B subunit of CT) in any of these lysates, including pJM17 . On the other hand, under non-dissociating conditions of electrophoresis, two polypeptides contained in cell lysates of pCI and 2B8, were detected on an immunoblot probed with anti-CT (Fig . 6, lanes 5 and 6, respectively) . One of them migrated in a manner similar to that of the A subunit of cholera toxin and the other to the whole CT (Fig . 6, lane 4) . However, when affinity-purified antibodies to CTB were used, only one band was detected in the cell lysates of pC1 and i.B8 which corresponded in size to the whole CT (data not shown) . In vitro DNA-directed transcription/ translation . To identify the polypeptide(s) encoded by the Salmonella enterotoxin gene (stx), we performed in vitro DNA-directed transcription/translation of the recombinant plasmid pC1 and bacteriophage ),B8 . Figure 7(A) showed that the 5 .0 kb DNA fragment of Salmonella inserted in pC1 encoded at least 8-10 more polypeptides . Among these proteins, two polypeptides of molecular size 30 and 11 kDa [Fig . 7(A), lane 2] corresponded to the size of proteins specified by the cloned ctx gene in pJM17 [Fig . 7(A) and (B), lane 3] . However, these specific polypeptides were not detected in the DNA of pBR322 alone [Fig . 7(A), lane 1] . The 30 kDa protein, which migrated as a doublet in pC1, was masked partially by a heavy protein band of (3-lactamase [Fig . 7(A), lanes 1 and 2] encoded by pBR322 . These two unique protein bands of 30 and 11 kDa were identified also in the recombinant phage 2B8 [Fig . 7(B), lane 2], but not in the ). phage control [Fig . 7(B), lane 1 ] . These data suggested that both pC1 and ).B8 carried DNA sequences

320

R . Prasad et al.

Fig . 3 . Effect of the cloned Salmonella enterotoxin on the morphology of cultured CHO cells . The G N„ ELISA-positive fractions from P-11 column were pooled and dialysed against 10 mm sodium phosphate buffer (pH 7 .0) . Equal volumes of the partially purified enterotoxin preparations were mixed with either preimmune (A) or immune serum to CT (B) and incubated for 1 h at 37'C prior to examining their effect on CHO cells . (C) The effect of heating (60°C, 5 min) on the enterotoxic activity of the Salmonella enterotoxin . (D) Control untreated CHO cells . The cells were fixed and stained with Giemsa's stain . Bar represents 50 um .

similar to pJM17 and possibly encoded polypeptides analogous to the un-nicked A subunit and monomeric /l-chain of CT .

Discussion Bacterial enterotoxins have been shown to be important determinants in the pathogenic mechanisms of several diarrheal diseases ."' However, little is known about the nature of Salmonella enterotoxin . Our laboratory and other investigators have demonstrated cholera toxin-like enterotoxin antigen and biological activity in various isolates of ."' At the DNA level, sequence homology has been surmised between Salmonella Salmonella chromosomal DNA and the cholera toxin gene 21,2' as well as with the E. coli LT1 gene . 26 To study the nature of the Salmonella enterotoxin, cell lysates from two E. coli clones (pC1 and 2B8) carrying a DNA fragment from S . typhimurium Q1, which reacted with the ctx gene probe, were examined for biological activity as well as for molecular and antigenic structure of the toxin molecule . It is clear from the data shown in Table 1 that the cell lysates of our bacteriophage



CT-like Salmonella enterotoxin

321

600 (A)

I 500

E

400

I a. a 2

200

U T U

I 100

I

0

I

1

(B)

I

2500

2000 rn

1500 N W

a 1000 I

500

0

I

I

PBS

I

a

xB8

PCI

CT

Fig . 4. Cyclic AMP and prostaglandin levels in CHO cells exposed to the cloned Salmonella enterotoxin preparations . The CHO cell monolayers were treated with the cell lysates of E. coli VCS257[EMBL3] (2), E. coli VCS257[1B8] (2B8), biologically active pool of P-11 phosphocellulose chromatographed cell lysates of E. coli DH5a[pCl] [pC1], cholera toxin (CT), and phosphate-buffered saline (PBS) . (A) cAMP (pmoles/mg) ; (B) Levels of PGE2 (pg/mg) in various preparations . The vertical axes represent PGE 2 and cAMP per mg of CHO cells protein . Results were expressed as the mean±SD of duplicate cAMP and PGE2 determinations on replicate cell monolayers . The data was analysed statistically by an unpaired, two-tailed Student's t-test .

clone [),B8] elicited a positive enterotoxic response in rabbit intestinal loops . The plasmid clone [pC1 ] failed to show enterotoxic activity in the rabbit intestinal loop model . However, the cell lysates of the pC1 clone caused elongation of CHO cells which was indistinguishable from the one caused by the cell lysates of clone AB8 and cholera toxin (Figs 1 and 3), although it exhibited lower titers on CHO cells . Positive enterotoxic activity in rabbit intestinal loops by M .B8 clone and not by pC1 clone suggests that the 14 .5 kb fragment from the Salmonella chromosomal DNA which was inserted in ,,B8 may contain a trans-activating region . This region may be important for the in vivo regulation of expression of toxin gene . The trans-acting element may not be present in the cloned DNA fragment of pC1 . The presence of such a trans-

R . Prasad et al .

322

3 4 92.5 69

5

14.3

Fig . 5 . Immunoblot analysis of the cloned Salmonella enterotoxin . Bacterial whole cells (from 1 .5 ml of culture) were lysed in 100 µl sample buffer before electrophoresis, 41 or cell lysates prepared using a French press . All cell lysates were electrophoresed under reducing conditions on 12% SDS-PAGE gels . The separated proteins were transferred onto a nitrocellulose membrane and probed with the affinity-purified cholera antitoxin . Lysed whole cells (30 µl) or 5-pi of the cell lysates were loaded onto the gel . Lanes : 1, lysed whole cells of E. coli DH5a[pBR322] ; 2, cell lysates of E. co/i DH5a[pBR322] ; 3, cell lysates of E. coli VCS257[EMBL3] ; 4, rainbow molecular mass markers ; 5, pure cholera toxin (1 µg) ; 6, lysed whole cells of E. coli DH5a[pCl ] ; 7, cell lysates of E. coli DH5a[pC1 ] ; 8, cell lysates of E. coli VCS257[).B8] ; and 9, lysed whole cells of E. coil MS371 [pJM17] . Numbers on the left indicate molecular masses of the protein markers (kDa) . The 21 .5 kDa band in lane 5 represents nicked CTA .

activating regulatory gene (ToxR) which controls the regulation of CT in V. cholerae has been demonstrated earlier by Miller et al. 31 A similar trans-acting gene whose product positively regulates the synthesis of exotoxin A in Pseudomonas aeruginosa has also been reported . 32 Additional similarities between the cloned Salmonella enterotoxin and CT were observed by the neutralization studies . The CHO cell elongation activity of the pC1 and 2B8 clones, like that of CT, was neutralized completely by monospecific antisera directed against CT [Fig . 3(B)], and CTB but not by anti-CTA (data not shown) . Furthermore, G M1 -ganglioside also completely neutralized the CHO cell elongation activity associated with the cloned Salmonella enterotoxin (data not shown), suggesting the presence of similar receptors on CHO cells for CT and the cloned Salmonella enterotoxin . We observed also that the enterotoxic response caused by cell lysates of the ),B8 clone in rabbit ileal loops was neutralized by antibodies to CT (Table 1) . Like CT, the Salmonella enterotoxin also was heat labile [Fig . 3(C)] . These observations were consistent with the characteristics of a native Salmonella enterotoxin reported earlier. 13 However, unlike CT produced by E. coli MS371 [pJM17], 28 the specific activity of the cloned Salmonella enterotoxin was not increased by trypsin treatment . These data suggest that the CTA-like subunit of the cloned Salmonella enterotoxin does not require additional proteolytic processing to express full enterotoxic activity .



CT-like Salmonella enterotoxin

323

Fig . 6 . Immunoblot analysis of the cloned Salmonella enterotoxin . The cell lysates (5 µl) were electrophoresed under non-dissociating conditions (without SDS, 2-ME and boiling) on a 12% PAGE gels . The separated proteins were transferred onto a nitrocellulose membrane and probed with affinity-purified cholera antitoxin . Lanes : 1, cell lysates of E. coli VCS257[EMBL3] ; 2, cell lysates of E. coli DH5a[pBR322] ; 3, rainbow molecular mass markers (not resolved under non-dissociating conditions) ; 4, pure cholera toxin (1 pg) ; 5, cell lysates of E. coli DH5a[pCl] ; 6, cell lysates of E. coli VCS257[%B8] .

Contrary to some of our observations with the cloned Salmonella enterotoxin, Wallis reported earlier that polymyxin B extracts of various Salmonella isolates which caused fluid accumulation in the rabbit-ileal loops, were inactive in the in vitro assays such as CHO cells, Y-1 adrenal and Vero cells, as well as an in vivo infant mouse assay . Furthermore, Clarke et al. 20 demonstrated a CT-related antigen in several strains of Salmonella using fluorescent antibodies to CT as well as by immunoblots but they could not detect this CT-related antigen by ELISA when G M1 -ganglioside or anti-CT was used as an anchor . Recently, an enterotoxin gene that has some similarities with CT and LT1, has been cloned from S . typhi using a DNA probe made to the B subunit of LT1 . 26 This enterotoxin caused fluid secretion in rat ileal loops, elongated CHO cells, and was detected in an ELISA using anti-LT1 . However, the toxin did not bind to G M1 -ganglioside . 26 These data suggest that although Salmonella enterotoxin has cross-reactivity with CT and LT1, there are clearly some unique features in the family of these enterotoxins . The role of cAMP in mediating diarrheal diseases by cholera toxin and E. coli heat-labile enterotoxin is very well documented .", " Our laboratory demonstrated previously29 30 that crude lysates of Salmonella elevated the levels of cAMP in intestinal epithelial cells and CHO cells in vitro, as well as during experimental salmonellosis . To confirm our previous findings, we examined the effect of cloned Salmonella enterotoxin on cAMP levels in CHO cells . Our data clearly demonstrated (Fig . 4) that the cloned Salmonella enterotoxin, like CT, increased intracellular levels of cAMP, which presumably stimulated fluid and electrolyte secretion . The possible involvement of prostaglandins in the pathogenesis of cholera has been suggested . 35 Recent observations from our laboratory have indicated that intestinal

et al."



R . Prasad et al.

324 (A) I

2

(B) 3

I

2

3

Fig . 7 . In vitro expression of the cloned Salmonella enterotoxin gene . (A) [ 35S] methionine-labeled plasmid-specific proteins encoded by pBR322 (lane 1), by pC1 (lane 2) and by pJM1 7 (lane 3) . (B) [ 35 S] methionine-labeled bacteriophage-specific proteins encoded by A phage (lane 1), by AB8 (lane 2), and by plasmid pJM17 (lane 3) . Arrows indicate the positions of 30 and 11 kDa polypeptides . The dash indicated the positions of the protein markers (for molecular masses, see Fig . 5) .

fluid secretion by CT correlates better with the prostaglandin content of the intestinal fluids than with the tissue concentration of cAMP ." We have also reported that challenge of intestinal loops with S . typhimurium or isolated intestinal cells with crude lysates of S . typhimurium causes an increase in cellular PGE 2 ." These data prompted us to examine whether or not the cloned Salmonella enterotoxin elevated PGE 2 release from CHO cells . Our data indicated that lysates of the 2B8 clone and P-11 pool of the pC1 preparations indeed increased PGE 2 release from CHO cells (Fig . 4) significantly above the respective vector controls, in a manner analogous to CT ." To study the structure and antigenic nature of the toxin molecule, we performed an in vitro transcription/translation analysis of isolated DNA from the clones pC1 and )B8 . Both of these clones exhibited the presence of at least two unique polypeptides (30 and 11 kDa), which were identical in size to the un-nicked A and B subunits of CT synthesized by pJ M17 . 28 Similar to our observations, Finkelstein et al . 23 had shown that the native enterotoxin isolated from S . typhimurium W118-2 contained A and B subunit structures like CT and LT1 . It remains to be determined whether the Salmonella enterotoxin also contains other peptides or epitopes [Fig . 7(A) and (B)] that are unrelated to CT . The cross-reactivity of CT with the Salmonella antigen has been reported by several groups .' 3-15,20,21,23 Our immunoblot analysis of crude lysates of E. coli DH5a[pC1 ], E. coli VCS257[2B8], and a P-11 phosphocelIulose chromatographed preparation of E

CT-like Salmonella enterotoxin

325

co/i DH5a[pC1 ] exhibited a CTA-cross-reactive polypeptide with a molecular mass of 30 kDa (Figs 2 and 5) . This 30 kDa polypeptide corresponded to the un-nicked A subunit of CT in the whole cell lysate of pJ M11 . Pearson and Mekalanos 28 demonstrated clearly that the E. coli cells were defective in proteolytic processing of the A subunit of CT encoded by pJM17 . Cross-reactivity of the 30 kDa protein band (putative A subunit of the cloned Salmonella enterotoxin) with antibodies to CTA suggested some similarity in the primary amino acid sequences of the A subunit of these two toxins . At this time, it is unclear why our antibodies, which react with pure CTB, failed to detect the 11 kDa protein band in immunoblot analyses of lysates of pC1, , B8, and pJM17 (Fig . 5) . We must remember, however, that the 11 kDa B subunit is prepared from CT that has been secreted from the bacterial cells . I n contrast, we examined lysates without post-translational processing (secretion) . Nevertheless, indirect evidence from neutralization experiments, G M ,-ELISA, fluid secretion in rabbit intestinal loops, and CHO cell elongation, suggested that the cloned Salmonella enterotoxin could have a functional CTB-like subunit . In conclusion, we have cloned the genetic locus from S. typhimurium Q1, which encodes a cholera toxin-like enterotoxin . The evidence presented in this report suggests an evolutionary relatedness between enterotoxins of S . typhimurium and V. cholerae . The precise role of this toxin in the pathogenesis of salmonellosis will be established only with additional genetic analysis and when this toxin has been purified to homogeneity .

Materials and methods

Bacterial strains and plasmids . The strains of S. typhimurium and E. coli were stored at -70'C in Luria-broth (LB) containing 25% glycerol (v/v) . Competent cells of E. coli DH5a (BRL, Gaithersburg, Maryland) were used in the transformation experiments and E. co/i VCS257 cells (Stratagene, La Jolla, California) were used to infect and screen the phage library . A strain of E. co/i MS371 harboring plasmid pJM17 (carrying the ctx gene) was kindly provided to us by Dr J . J . Mekalanos, Harvard Medical School, Boston, Massachusetts . DNA manipulations . All of the DNA manipulations were performed by standard procedures .11,31 A synthetic oligonucleotide made to a region of the B subunit of cholera toxin gene (position 1625-1645) was employed to detect the Salmonella enterotoxin gene (stx) which was subsequently cloned into pBR322, as well as into bacteriophage EMBL3 . For cloning into pBR322, we used an EcoRl-digested chromosomal DNA fragment of S . typhimurium Q1 that reacted with the above mentioned probe in a Southern blot analysis . A genomic library of S . typhimurium Q1 was constructed by partially digesting the DNA with Sau3A restriction enzyme to produce fragments of 10 to 20 kb . These fragments were ligated into BamHl-digested IEMBL3 arms (Promega, Madison, Wisconsin) . The gene library then was packaged in vitro using Gigapack II gold packaging extract (Stratagene) and the recombinant phage was transduced into E. coli VCS257, according to the instructions of the manufacturer . Finally, pC1 and )B8 clones were selected from the plasmid and phage libraries, respectively, for further studies .

Bacterial cultivation and cell lysate preparation . Eight Fernbach flasks, each containing 500 ml of L-broth (with 15 pg/ml tetracycline), were inoculated with 1 ml E. co/i DH5a[pC1 ] grown overnight in L-broth . The bacterial cultures were incubated at 37 ° C on a rotary shaker (150 rpm) for 16-18 h . The cells were harvested by centrifugation (10000xg for 15 min) in a refrigerated centrifuge (Sorvall RC5B) . However, the E. co/i VCS257 cells infected with the bacteriophage carrying the stx gene [2B8] were grown in 1-2 1 of NZYCM medium (N-Z Amine A, 10 g/I, NaCl, 5 g/l, yeast extract, 5 g/l, casamino acids 1 g/l, and MgSO, • 7H 2 0, 2 g/l, pH 7 .5) until the density of the culture at 600 nm reached to 0 .5 . Usually 1 x 10 10 cells of E. coli were infected with 5x 10' bacteriophages . The cell pellet was washed once with 10 mm sodium phosphate buffer (pH 7 .0) containing 0 .2 mm phenylmethylsulfonyl fluoride and resuspended



326

R . Prasad et al

in the same buffer to 1 /50th the original volume of the culture . The cells were then disrupted in a French Pressure cell press (AMINCO) at a pressure of 18000 psi . The cell lysatas were centrifuged at 27000xg for 30 min to remove the cell debris and the supernatants were then ultracentrifuged at 120000xg for 2 h to remove the cell membranes . The clear supernatants were dialysed in the above mentioned buffer before examining for enterotoxic activity or further purification on a column of phosphocellulose P-11 (Whatman, Clifton, New Jersey) . The biological activity of the enterotoxin was detected in vivo by the rabbit intestinal loop assay and in vitro by CHO cell elongation assay . Rabbit intestinal loop assay . The biological activity of the cloned stx gene product(s) was assayed in young New Zealand white rabbits, according to the method of De and Chatterje . 38 Briefly, 1 .0 ml of each cell lysate of E. co/i DH5a[pC1] and E. coli VCS257[ B8] was injected into a 10-cm ligated rabbit intestinal loop . Phosphate-buffered saline and cell lysates of E. coli DH5a[pBR322] and E. coil VCS257[AEMBL3] were used as negative controls . Following overnight incubation, rabbits were sacrificed and examined for fluid accumulation . CHO cell elongation assay . Chinese hamster ovary cells from ATCC (Rockville, Maryland) were grown in 75-cm 2 sterile tissue culture flasks at 37°C with 5% CO 2 in Ham's F-12 medium (Gibco Laboratories, Gaithersburg, Maryland) supplemented with 10% fetal calf serum (FCS), 100 U/ml penicillin G, 100 pg/ml streptomycin, 50 pg/ml gentamicin, and 50 U/ml mycostatin . For the CHO cell elongation assay, 39 the cells were suspended in Ham's F-12 medium containing 1 % FCS, 100 U/ml penicillin G, 100 pg/ml streptomycin, 50 pg/ml gentamicin, and 50 U/ml mycostatin . Aliquots (200 pl) of the CHO cell suspension, containing approximately 2000 cells, were delivered to each well of 96-well microtiter plates . After 3-4 h of incubation at 37°C with 5% CO 2 , 25-µl volumes of the two-fold diluted toxin preparations were added to the wells . Incubation of the plates was continued further under the same conditions for 24-36 h . After the incubation period, the CHO cells were fixed with 70% methanol for 15 min, stained with Giemsa's stain, and observed for the elongation of cells . Measurement of cyclic AMP . Cyclic AMP responses of CHO cells treated with the cloned Salmonella enterotoxin were assayed as described previously ." Measurement of prostaglandins . The prostaglandin response of CHO cells to Salmonella enterotoxin was measured using a prostaglandin E 2 ( 3 H) radioimmunoassay (RIA) kit (Advanced Magnetics, Cambridge, Massachusetts) .` Preparation of antibodies to cholera toxin . Antiserum against pure cholera toxin was raised by immunization of rabbits as described previously ." Cholera toxin was purchased from List Biological Laboratories, Campbell, California, and CT subunits A and B were prepared by the HPLC method of Pearson et al ." Specific antibodies to CTA and CTB subunits were purified by affinity HPLC (Waters, Milford, Massachusetts), using Hydropore-EP mini columns (Rainin, Woburn, Massachusetts) coupled with CTA or CTB . Derivatization and subsequent elution of the antibodies from the columns were conducted as suggested by the manufacturer. Antigenic and biological characterization of the cloned Salmonella enterotoxin was performed using a mixture (1 :1) of affinity-purified antibodies to CTA and CTB . This mixture of antibodies was referred to as anti-CT and diluted 1 : 20 for immunoblot analyses as well as in the ELISA . Enzyme-linked immunosorbent assay (ELISA) . Polystyrene microtiter plates were coated with 100-µl of a solution (10 pg/ml) of monosialogangliosides G M , ( Sigma, St Louis, Missouri) in a carbonate buffer (Na 2 CO 3 1 .59 g/l, NaHCO 3 2 .9 g/I ; pH 9 .6) . The plates were incubated in a humid chamber overnight at room temperature . The next morning, all wells were washed three times with phosphate-buffered saline (PBS, pH 7 .4) containing 0 .05% Tween-20 and 0 .1% bovine serum albumin (PBS-T-BSA) . Two-fold toxin dilutions in PBS-T-BSA buffer were added to the wells and the plates were incubated for 2 h at room temperature . The wells were washed three times with PBS-T-BSA buffer and then the affinity-purified polyclonal antibodies to CT were added to each well and the plates were incubated for a further 2 h at room temperature . The plates were washed again with PBS-T-BSA as mentioned above, and goat anti-rabbit immunoglobuln (diluted 1 :3000 in PBS-T-BSA) conjugated with alkaline phosphatase (BioRad, Richmond, California) was added to each well for 2 h at room temperature . After washing three times with PBS-T-BSA, p-nitrophenyl phosphate (1 mg/ml, Sigma, St Louis, Missouri) in carbonate buffer (pH 9 .6) was added to each well . The reaction was stopped



CT-like Salmonella enterotoxin

327

after 60 min by adding 25-pI of 3M NaOH to each well and the intensity of yellow color was measured spectrophotometrically at 405 nm . The purified CT was used as a positive control in the ELISA .

P-11 Phosphocellulose chromatography . Chromatography of whole cell lysates of E. co/i DH5x[pC1] was performed on a column (1 .5x30 cm) of P-11 phosphocellulose (Whatman) equilibrated with 10 mm sodium phosphate buffer (pH 7 .0), as described previously ." The bound material from the column was then eluted with 200 mm sodium phosphate buffer (pH 7 .4) . The eluted fractions (4 ml each) were examined for the presence of CT-like antigen in the G M ,-ELISA using antibodies to CT . The fractions having CT-like antigen were pooled, dialysed against 10 mm sodium phosphate buffer (pH 7 .0) and used for immunological and biological characterization of the antigen .

Neutralization of the cloned Salmonella toxin . Partially purified enterotoxin preparations (P11) or lysates of pC1 and ).B8 were mixed in a 1 :1 ratio with preimmune or immune antitoxic serum to CT and incubated for 1 h at 37°C . Two-fold dilutions of each of these samples (25 pl) were added to the CHO cell wells . After 36 h of incubation, the CHO cells were examined microscopically for elongation . The treated lysates from E. co/i VCS257[)B8] were tested also in the rabbit intestinal loop assay to demonstrate whether or not antibodies to CT could neutralize the biological activity associated with the cloned Salmonella enterotoxin .

SDS-PAGE and immunoblot analysis . SDS-PAGE was performed according to the method of Laemmli 43 using 12% discontinuous slab gels . The samples were boiled for 4 min in the presence of 2% SDS and 5% 2-mercaptoethanol . The proteins were transferred onto a nitrocellulose membrane (BioRad) in a transblot apparatus (BioRad), according to the manufacturer's instructions . Subsequently, the nitrocellulose membranes were probed with the affinity-purified antibodies to CT . In a non-dissociating PAGE analysis, SDS and 2mercaptoethanol were omitted and the samples were not boiled . Rainbow molecular weight markers (Amersham, Arlington Heights, Illinois) were used as standards .

In vitro protein synthesis. In vitro transcription/translation of the cloned Salmonella enterotoxin gene was performed utilizing isolated DNA of our clones pC1 and i.B8 . A prokaryotic DNAdirected translation kit was purchased from Amersham, and used according to the manufacturer's recommendations. Subsequently, 35 S-methionine-labeled proteins were separated on a 12% discontinuous SDS-PAGE slab gel . The resolved polypeptides were visualized by fluorography using EN 3 HANCE (DuPont, Massachusetts) .

Statistical analysis . The cyclic AMP and PGE 2 data were analysed by an unpaired, two-tailed Student's t-test .

This work was supported by research grant R01 All 8401 from the National Institutes of Health . The authors thank M . Susman for editorial assistance during manuscript preparation and C . Jackson for performing the cAMP and PGE 2 assays .

References 1 . Chikami GK, Fierer J, Guiney DG . Plasmid mediated virulence in Salmonella dublin demonstrated by the use of a Tn5-oriT construct . Infect Immun 1985 ; 50 : 420-4 . 2 . Christine MW, Baird GD, Manning EJ . A common virulence region on plasmids from eleven serotypes of Salmonella . J Gen Microbiol 1988 ; 134 : 975-82 . 3 . Gulig PA, Curtiss III R . Plasmid associated virulence of Salmonella typhimurium . Infect Immun 1987 ; 55 :2891-901 . 4 . Jones GW, Robert DK, Svinarich DM, Whitfield HJ . Association of adhesive, invasive and virulent phenotypes of Salmonella typhimurium with autonomous 60-megadalton plasmids . Infect Immun 1982;38 :476-86 .



328

R . Prasad et al.

5 . Nakamura M, Sato S, Ohya T, Suzuki S, Ikeda S . Possible relationship of a 36-megadelton Salmonella enteritidis plasmid to virulence in mice . Infect Immun 1985 ; 47 : 831-3 . 6 . Galan JE, Curtiss III R . Cloning and molecular characterization of genes whose products allow Salmonella typhimurium to penetrate tissue culture cells . Proc Natl Acad Sci USA 1989; 86 : 6383--7 . 7 . Fields PI, Groisman EA, Heffron F . A Salmonella locus that controls resistance to microbial proteins from phagocytic cells . Science 1989 ; 243 : 1059-62 . 8 . Law D . Virulence factors of enteropathogenic Escherichia co/i. J Med Microbiol 1988 ; 26 : 1-10 . 9 . Finkelstein RA . Cholera . Crit Rev Microbiol 1973 ; 2 : 553-623 . 10 . Dorner F . Escherichia co/i enterotoxin . Purification and partial characterization . In : Schlessinger D, ed . Microbiology . Washington, DC : American Society 1975 ; 242 . 11 . Koupal LR, Deibel RH . Assay, characterization and localization of an enterotoxin produced by Salmonella . Infect Immun 1975 ; 11 : 14-22 . 12 . Sandefur PD, Peterson JW . Isolation of skin permeability factors from culture filtrates of Salmonella typhimurium. Infect Immun 1976; 14 : 671-9 . 13 . Sandefur PD, Peterson JW . Neutralization of Salmonella toxin-induced elongation of Chinese hamster ovary cells by cholera antitoxin . Infect Immun 1977 ; 15 : 988-92 . 14 . Peterson JW . Salmonella toxin . Pharmacol Ther 1980 ; 11 : 719-24 . 15 . Houston CW, Koo FCW ; Peterson JW . Characterization of Salmonella toxin released by mitomycin Ctreated cells . Infect Immun 1981 ; 32 : 916-26 . 16 . Singh SP, Sharma VD, Singh IP . Isolation of an enterotoxin factor from Salmonella saintpaul. FEMS Microbiol Lett 1985; 26 : 301-4 . 17 . Wallis TS, Starkey WG, Stephen J, Haddon SJ, Osborne MP, Candy DCA . Enterotoxin production by Salmonella typhimurium strains of different virulence . J Med Microbiol 1986 ; 19-23 . 18 . Sedlock DM, Koupal LR, Deibel RH . Production and partial purification of Salmonella enterotoxin . Infect Immun 1978 ; 20 : 375-80 . 19 . Peterson JW, Sandefur PD . Evidence of a role for permeability factors in the pathogenesis of salmonellosis . Am J Clin Nutr 1979 ; 32 : 197-209 . 20 . Clarke GJ, Q1 G-M, Wallis TS et al. Expression of an antigen in strains of Salmonella typhimurium which reacts with antibodies to cholera toxin . J Med Microbiol 1988; 25 : 139-46 . 21 . Q1 G-M, Clarke GJ, Wallis TS, Stephen J . The influence of cultural conditions on the expression in Salmonella typhimurium of an antigen related to cholera toxin . J Med Microbiol 1989; 30 : 213-17 . 22 . Gill DM, Clements JD, Robertson DC, Finkelstein RA. Subunit number and arrangement in Escherichia co/i heat-labile enterotoxin . Infect Immun 1981 ; 33 : 677-82 . 23 . Finkelstein RA, Machlewiz BA, McDonald RJ, Finkelstein MB . Isolation and characterization of a cholera-related enterotoxin from Salmonella typhimurium . FEMS Microbiol Lett 1983 ; 17 : 239-41 . 24 . Chopra AK, Houston CW, Peterson JW, Mekalanos JJ . Chromosomal DNA contains the gene for Salmonella enterotoxin . FEMS Microbiol Lett 1987 ; 43 : 345-9 . 25 . Chopra AK, Houston CW, Peterson JW, Prasad R, Mekalanos JJ . Cloning and expression of Salmonella enterotoxin gene . J Bacteriol 1987 ; 169 : 5095-100 . 26 . Fernandez M, Sierra-Madero J, de la Vega H et al. Molecular cloning of a Salmonella typhi LT-like enterotoxin gene . Mol Microbiol 1988; 2 : 821-6 . 27 . Maniatis T, Fritsch EF, Sambrook J . Molecular cloning : a laboratory manual . Cold Spring Harbor : Cold Spring Harbor Laboratory Press, 1982 . 28 . Pearson GD, Mekalanos JJ . Molecular cloning of Vibrio cholerae enterotoxin genes in Escherichia coli K-12 . Proc Nat] Acad Sci USA 1982; 79 : 2976-80 . 29 . Peterson JW, Molina NC, Houston CW, Fader RC . Elevated cAMP levels in intestinal cells during experimental cholera and salmonellosis . Toxicon 1983 ; 21 : 761-75 . 30 . Duebbert IE, Peterson JW. Enterotoxin-induced fluid accumulation during experimental salmonellosis and cholera : involvement of prostaglandins synthesis by intestinal cells . Toxicon 1985 ; 23 : 157-72 . 31 . Miller VL, Taylor RK, Mekalanos JJ . Cholera toxin transcriptional activator ToxR is a transmembrane DNA binding protein . Cell 1987; 48 : 271-9 . 32 . Hamood AN, Iglewski BH . Expression of the Pseudomonas aeruginoas ToxA positive regulatory gene (regA) in Escherichia coli. J Bacteriol 1990 ; 172 : 589-94 . 33 . Field M . Regulation of small intestinal ion transport by cyclic nucleotides and calcium . In : Field M, Fordtran JS, Schultz SG, eds . Secretory diarrhea . Baltimore : Waverly Press, 1980 ; 21 . 34 . Finkelstein RA . Progress in the study of cholera and related enterotoxins . In : Bernheimer AW, ed . Mechanisms in bacterial toxicology . New York : John Wiley 1976 ; 54 . 35 . Bennett A . Cholera and prostaglandins . Nature 1971 ; 231 . 536 . 36 . Peterson JW, Ochoa LG . Role of prostaglandins and cAMP in the secretory effects of cholera toxin . Science 1989 ; 245 : 857-9 . 37 . Ausubel FM, Brent R, Kingston RE et al., eds . Current protocols in molecular biology . New York : John Wiley, 1989 . 38 . De SN, Chatterje DN . An experimental study of the mechanism of action of Vibrio cholera on the intestinal mucous membrane . J Pathol Bacteriol 1953; 66 : 559-62 . 39 . Guerrant RL, Brunton LL, Schnaitman TC, Rebhum LI, Gilman AG . Cyclic adenosine monophosphate and alteration of Chinese hamster ovary cell morphology : a rapid, sensitive in vitro assay for the enterotoxins of Vibrio cholerae and Escherichia coli. Infect Immun 1974; 10 : 320-7 .



CT-like Salmonella enterotoxin 40 .

329

Hejtmancik KE, Peterson JW, Markel DE, Kurosky A . Radioimmunoassay for the antigenic determinants of cholera toxin and its components . Infect Immun 1977 ; 17 : 621-8 . 41 . Pearson SD, Dixon J D, Nothwehr SF, Kurosky A . Isolation of high-specific-activity subunits of cholera toxin by reversed-phase high-performance liquid chromatography . J Chromatogr 1986 ; 359 : 413-421 . 42 . Mekalanos JJ, Collier RJ, Romig WR . Purification of cholera toxin and its subunits : new methods of preparation and the use of hypertoxinogenic mutants . Infect Immun 1978; 20 : 552-8 . 43 . Laemmli UK . Cleavage of structural proteins during the assembly of the head of bacteriophage T4 . Nature 1970 ; 227 : 680-5 .