Purification and preliminary characterization of the zonula occludens ...

FEMS Microbiology Letters 194 (2001) 1^5

www.fems-microbiology.org

Puri¢cation and preliminary characterization of the zonula occludens toxin receptor from human (CaCo2) and murine (IEC6) intestinal cell lines Sergio Uzzau a

a;

*, Ruliang Lu a , Wenle Wang a , Cara Fiore a , Alessio Fasano

a;b

Division of Pediatric Gastroenterology and Nutrition and Gastrointestinal Pathophysiology Section, Center for Vaccine Development, University of Maryland, School of Medicine, Baltimore, MD 21201, USA b Department of Physiology, University of Maryland, School of Medicine, Baltimore, MD 21201, USA Received 3 July 2000; accepted 30 September 2000

Abstract In the present study, we report the preliminary characterization of the epithelial cell receptor for Vibrio cholerae zonula occludens toxin (Zot). Zot receptor was purified by ligand-affinity chromatography. Analysis of affinity-purified preparations by polyacrylamide gel electrophoresis revealed a protein of ca. 66 kDa. Partial N-terminal sequence obtained from purified murine and human Zot receptor revealed homology between the two proteins and with human K-1-chimaerin. Zot protein domain(s) involved in receptor binding were also analyzed by constructing several in frame deletion derivatives of a recombinant fusion Zot protein tagged with maltose binding protein. Our results suggest that Zot binding to its cellular membrane receptor requires a sequence that spans between amino acids 118 and 299. ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Vibrio cholerae; Zonula occludens toxin; Tight junction; Receptor

1. Introduction Tight junctions (tj) are one of the hallmarks of absorptive and secretory epithelia. As a barrier between apical and basolateral compartments, they selectively control the passive di¡usion of ions and small water-soluble solutes through the paracellular pathway, thereby counter-regulating any gradients generated by transcellular pathways. The tj represents the major barrier within this paracellular pathway and the electrical resistance of epithelial tissues seems to depend on the number of transmembrane protein strands and their complexity [1]. Evidence now exists that tj, once regarded as static structures, are in fact dynamic, and readily adapt to a variety of developmental [2], physiological [3], and pathological [4] circumstances. These adaptive mechanisms are still incompletely understood. We have demonstrated that zonula occludens toxin (Zot), a protein elaborated by Vibrio cholerae, reversibly increases the permeability of tj [5]. Zot action is mediated * Corresponding author. Present address: Dipartimento di Scienze Biomediche, Universita© degli Studi di Sassari, Viale S. Pietro 43B, 07100 Sassari, Italy. Tel.: +39 (079) 228303; Fax: +39 (079) 212345; E-mail : [email protected]

by a cascade of intracellular events that lead to a protein kinase C (PKC) K-dependent polymerization of actin micro¢laments strategically localized to regulate the paracellular pathway [6]. It has been hypothesized that Zot possesses multiple domains that allow a dual function as a morphogenetic phage protein and as an enterotoxin [7]. Zot binding to the surface of rabbit intestinal epithelium has been shown to vary along the di¡erent regions of the intestine [8]. This binding distribution coincides with the regional e¡ect of Zot on intestinal permeability and with the preferential F-actin redistribution induced by Zot in the mature cells of the villi [6,8]. In the present study we aimed to purify and partially characterize Zot receptor from intestinal epithelial cell lines and to identify the Zot region required for receptor binding. 2. Materials and methods 2.1. Construction and puri¢cation of maltose binding protein (Mbp)^Zot in frame deleted mutants zot gene has been cloned in plasmid vector pMal-c (New England Biolabs) according to standard procedures (Table

0378-1097 / 01 / $20.00 ß 2001 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII: S 0 3 7 8 - 1 0 9 7 ( 0 0 ) 0 0 4 7 8 - X

FEMSLE 9697 21-12-00

2

S. Uzzau et al. / FEMS Microbiology Letters 194 (2001) 1^5

Table 1 Bacterial strains Strain

Relevant genotype

Reference

DH5K(pMal-c) DH5K(pSU201) DH5K(pSU202) DH5K(pSU203) DH5K(pSU204) DH5K(pSU206)

malE malE: :zot malE: :zot malE: :zot malE: :zot malE: :zot

[17] [8] This This This This

a

v1^99a v118^265a v118^299a v301^399a

study study study study

Deleted amino acid residues.

1). pMal-c carries the malE gene for Mbp and allows cloning and expression of Mbp fusion proteins. Inserts to be cloned as malE fusion genes were obtained by PCR with primers described in Table 2. For those clones with in frame deletions in the central part of the molecule (pSU203 and pSU204), pSU201 was digested with StuI and HindIII and zot 3P sequence was replaced with amplicons obtained with F7/F2 and F11/F2 primer pairs, respectively. Expression and puri¢cation of Mbp^Zot and its deleted derivatives were carried out as previously described [8]. The typical yield from 1 l of culture was 2^3 mg of protein. 2.2. Analysis of Zot binding to cell lines Binding experiments were performed with several epithelial cell lines, including IEC6 (rat, intestine), CaCo2 (human, intestine), T84 (human, intestine), MDCK (canine, kidney), and BPA endothelial cells (bovine, pulmonary artery). For immuno£uorescence analysis, con£uent monolayers (2.0U105 cells) on glass slides were incubated at di¡erent time intervals (5 min, 30 min, 60 min), and temperatures (4³C or 37³C) with either 5.0U1039 M Mbp^Zot or 5.0U1039 M Mbp negative control. The cells were then ¢xed with cold methanol, washed three times with phosphate-bu¡ered saline (PBS), blocked with 1.0% (w/v) bovine serum albumin, and then incubated for 1 h with rabbit polyclonal anti-Mbp antibodies (New England Biolabs). The samples were then washed with PBS, and incubated with goat anti-rabbit IgG-FITC conjugated antibodies for 30 min, followed by incubation with Evans blue for 10 min. Immunoblotting analysis of Zot binding was performed on monolayers grown on 30-mm polystyrene dishes. Following incubation of monolayers with Mbp^Zot and in frame deletion Mbp^Zot derivatives (0.2 WM each) for 15 min at 37³C or 3 h at 4³C, cells were washed 10 times with cold PBS, suspended and lysed. Cell lysates were resolved by SDS^PAGE, transferred to PVDF membranes, and probed with anti-Mbp antibodies. Mocktreated monolayers were used to determine cell counts and total protein concentrations. To establish the speci¢city of Zot binding, radiolabeled Mbp^Zot was used. Brie£y, Mbp^Zot was iodinated using the IODO-BEADS Iodination Reagent according to the manufacturer's in-

structions (Pierce, Rockford, IL, USA). Con£uent IEC6 monolayers were washed, suspended in ice-cold PBS containing protease inhibitors, and incubated for 3 h at 4³C with 1 Wg [125 I]Mbp^Zot. Mock-treated monolayers were used to determine cell counts and total protein concentrations. Following incubation with [125 I]Mbp^Zot, cells were washed, lysed, and subjected to SDS^PAGE. Finally, the gel was dried and autoradiographed. These experiments were performed in the absence or presence of either 10- or 50-fold molar excess of unlabeled Mbp^Zot. 2.3. Puri¢cation of Zot binding protein A Mbp^Zot a¤nity column was prepared by immobilizing overnight, at room temperature, 1.0 mg of puri¢ed Mbp^Zot to a pre-activated gel (Aminolink, Pierce). The column was washed with PBS, and then loaded with a crude cell lysate obtained using either 106 IEC6 cells or CaCo2 cells. After a 90 min incubation at room temperature, the column was washed ¢ve times with 14 ml of PBS, and the proteins which bound to Mbp^Zot were eluted from the column with 4.0 ml of a solution comprising 50 mM glycine (pH 2.5), 150 mM NaCl, and 0.1% (v/v) Triton X-100. The pH of the 1.0-ml eluted fractions was immediately neutralized with 1.0 N NaOH. Collected fractions were subjected to 6.0^15.0% (w/v) gradient SDS^ PAGE under reducing conditions. The resolved proteins were transferred to a nitrocellulose membrane and subjected to N-terminal sequencing using a Perkin-Elmer Applied Biosystems Apparatus Model 494. 3. Results 3.1. Screening of di¡erent cell lines for Zot binding properties When incubated with Mbp^Zot fusion protein at 4³C and at increasing time intervals, CaCo2 (Fig. 1A) and IEC6 intestinal epithelial cells as well as BPA endothelial cells (data not shown) displayed a number of £uorescent yellow particles on the cellular surface. When incubated at Table 2 Primers Primer

Sequence (5P^3P)a

Clone obtained

F2

CCCAAGCTTGGGTCAAAATATACT (C)

F3 F5 F7 F8 F11

TCCCCCGGGGGAATGAGTATCTTT (N) TCCCCCGGGGGAGCCACCAATTTA (N) TCAAAGGCTACTGCTGGG (N) CCCAAGCTTGGGTCAGACAAAACCATC (C) CCTGTTGGTGATGAGCGT (N)

pSU201, pSU202, pSU203, pSU204 pSU201, pSU206 pSU202 pSU203 pSU206 pSU204

a Primers were designed to anneal to the sequence corresponding to the N-terminal (N) or the C-terminal (C) part of the molecule. Restriction sites are in bold: F3 and F5 (SmaI), F2 and F8 (HindIII). Sequences corresponding to those annealing to zot are shown in italics.

FEMSLE 9697 21-12-00


3

Fig. 1. Immuno£uorescence analysis of Zot binding to epithelial cell monolayers. Mbp^Zot binding to CaCo2 cell monolayers was localized on the surface of the epithelial cells (arrowheads) after incubation at 4³C (A) and both on the surface and within CaCo2 cells after incubation at 37³C (B); no staining was observed in CaCo2 cells when incubated at 37³C with Mbp (negative control) up to 60 min (C).

37³C, Mbp^Zot immuno£uorescent particles were localized both on the surface and within CaCo2 cells (Fig. 1B), suggesting that after engagement to its surface receptor, Mbp^Zot is internalized. On the contrary, no staining was observed in either T84 or MDCK cells when incubated with Mbp^Zot up to 60 min at 4³C or at 37³C (data not shown). CaCo2 cells did not display £uorescence staining after incubation with the Mbp negative control either at 4³C or at 37³C (Fig. 1C). The cell speci¢city of Mbp^Zot binding was con¢rmed by immunoblotting analysis of cell lysates after incubation of monolayers with Mbp^Zot or Mbp for 15 min at 37³C or 3 h at 4³C. Mbp^Zot bound to IEC6, CaCo2, and BPA but not to T84 and MDCK cells both at 4³C (data not shown) and at 37³C (Fig. 2). Since monolayers were incubated with Mbp^Zot at 37³C (see Section 2), Mbp^Zot association with cell lysates might represent internalization of the protein following the binding on the cellular surface. Mbp was not detected in any of the di¡erent cell lines following incubation at 4³C (data not shown) or at 37³C (Fig. 2).

Fig. 2. Immunoblotting analysis of Mbp^Zot binding to di¡erent cell lines. Cell monolayers were harvested 15 min after incubation at 37³C with Mbp^Zot or Mbp. Equal amounts of protein of whole cell lysates were separated on a 12% SDS^PAGE and blotted on PVDF ¢lters. Mbp^Zot binding to IEC6 (lane 2), BPA (lane 3), and CaCo2 (lane 4), but not MDCK (lane 5) and T84 (lane 6), was detected by anti-Mbp monoclonal antibodies. Mbp did not bind to IEC6 (lane 8), BPA (lane 9), CaCo2 (lane 10), MDCK (lane 11), and T84 (lane 12). Lanes 1 and 7 contain aliquots of puri¢ed Mbp^Zot and Mbp, respectively.

3.2. Speci¢city of Zot binding To determine whether a speci¢c Zot binding site exists on intestinal cells, competitive experiments were conducted using [125 I]Mbp^Zot. As shown in Fig. 3, labeling of Mbp^Zot did not a¡ect the capability of toxin to bind to IEC6 cells at 4³C. This binding was competitively displaced by 10 and 50 molar excess of unlabeled Mbp^Zot, thereby establishing speci¢city of Mbp^Zot binding. 3.3. Identi¢cation of a Zot region required for binding to the epithelial cell Four di¡erent in frame deletion derivatives of zot gene were constructed and the deleted sequences were expressed and puri¢ed as Mbp fusion protein. Deleted derivatives lacked N-terminal, C-terminal, or central regions of Zot sequence with lengths between 99 and 180 amino acids (Fig. 4, left). In order to evaluate the Zot sequence domain(s) that speci¢cally binds to Zot cell receptor, IEC6

Fig. 3. Autoradiography (A) and densitometry analysis (B) of Mbp^Zot binding to IEC6 cells. Radiolabeled [125 I]Mbp^Zot binding to IEC6 monolayers (lane 1) was displaced by 10 molar (lane 2) and 50 molar (lane 3) excess of `cold' Mbp^Zot. Densitometry analysis of autoradiogram shows the extent of signal reduction on lanes 2 and 3 as compared to lane 1.

FEMSLE 9697 21-12-00

4


(Fig. 5). A comparison of their N-terminal sequences revealed similarity between the murine and the human Zot binding protein. Furthermore, a search in the GenBank database revealed a high similarity to the N-terminal region of the human K-1-chimaerin (46% identity, 69% similarity), a potential signal transduction molecule that contributes to modulate cytoskeletal organization [9] (Fig. 5). Taken together, these results suggest that Zot interacts with a surface receptor(s) that may be coupled to an intracellular e¡ector mechanism(s) that regulates tj function. 4. Discussion Fig. 4. Immunoblot analysis of Mbp^Zot mutants binding to IEC6 cell monolayers. IEC6 cells were harvested 15 min after incubation at 37³C with Mbp or with each of the four Mbp^Zot mutants carrying di¡erent in frame deletions. The lysates normalized on the basis of their protein values were loaded on the gel (B). Aliquots of the corresponding puri¢ed proteins were also loaded as molecular mass markers (I). Deleted sequences (on the left) are represented with discontinued lines.

cell monolayers were exposed to each of the four deleted recombinant Mbp^Zot proteins at 37³C for 15 min. The ability of each deleted Mbp^Zot derivative to bind and, eventually, be internalized into IEC6 was assessed by immunoblotting analysis of IEC6 cell lysates. Wild-type Mbp^Zot sequence and Mbp were used as positive and negative controls, respectively (Fig. 4, right). Mbp^Zot binding to IEC6 cells was not a¡ected by elimination of either the N-terminal fragment or the C-terminal fragment of Zot sequence (respectively the ¢rst and the last 100 amino acids). Conversely, the toxin binding to IEC6 cells was reduced after sequence deletion around the central part of the molecule, and completely impaired by the deletion of Zot central sequence from amino acid 118 to 299, suggesting that the domain(s) involved in Zot binding to the epithelial cell receptor may be located within this sequence (Fig. 4). 3.4. Puri¢cation of Zot receptor In order to purify the intestinal Zot binding protein(s), both human (CaCo2) and murine (IEC6) intestinal cell lysates were loaded on a Mbp^Zot a¤nity column. The eluted fractions obtained from both IEC6 and CaCo2 cells contained a single protein band with a Mr of 66 kDa as observed by SDS^PAGE under reducing conditions (data not shown). The Zot binding proteins from CaCo2 and IEC6 cells were each subjected to N-terminal sequencing

A number of recent studies have shown that several virulence factors elaborated by microorganisms are able to exploit speci¢c physiologic pathways on their eukaryotic counterpart, in order to accomplish a successful pathogenic lifestyle [10]. The discovery of one such virulence factor, Zot, and the characterization of its mechanism of action prompted our group to further analyze the cascade of cellular events that leads to the modulation of tj permeability [8]. Moreover, we have recently described an endogenous tj modulator, named zonulin, that is structurally and immunologically related to Zot [11]. Zot undergoes processing that yields a V33-kDa N-terminal fragment, probably involved in CTXx assembly, and a V12-kDa C-terminal region that is secreted and might be involved in tj disassembly [7]. Interestingly, the Zot fragment biologically active on tj and zonulin share an N-terminus sequence motif that appears to be involved in the zonulin binding to its target receptor [7,12]. Fluorescence staining of Zot binding to the intestinal epithelium is maximal on the surface of the mature absorptive enterocytes at the tip of the villi and completely disappears along the surface of the immature crypt cells, suggesting that the expression of Zot/zonulin receptor is up-regulated during enterocyte di¡erentiation [8]. In the present study, we tested di¡erent cell lines as source for receptor characterization and puri¢cation. Since Zot binds to mature enterocytes but not to the less mature crypt cells, we investigated the presence of receptor(s) in several cell lines, including CaCo2 and T84 cells. We have shown that human intestinal epithelial CaCo2 (that resemble the mature absorptive enteric cell of the villi), but not T84 (crypt-like cells), express receptor(s) on their surface. The paucity of Zot binding in the crypt area may also re£ect the fact that this region is already leaky as compared to the more mature epithelium of the tip of the villi [13] and

Fig. 5. N-terminal sequences of Zot receptors puri¢ed from IEC6 (murine) and CaCo2 (human) cells and alignment with K-1-chimaerin.

FEMSLE 9697 21-12-00


thus might not need to express a signi¢cant amount of Zot/zonulin receptor. Also, competition experiments with radiolabeled Zot clearly demonstrated the speci¢city of the toxin binding to the epithelial monolayers. A¤nity chromatography allowed us to purify the Zot/ zonulin epithelial cell receptor. The human and murine receptors presented important sequence similarities between each other and with the human K-1-chimaerin, a neuron-speci¢c GTPase-activating protein for p21 rac, member of the Ras-related rho subfamily [9,14]. This small GTP binding protein is involved in the regulation of actin ¢laments and cytoskeletal organization in mammalian cells in response to several stimuli [15]. Of note, we have previously demonstrated that Zot induces a PKCK-dependent polymerization of actin ¢laments strategically located to modulate intercellular tj [6]. A possible explanation for these ¢ndings is that the 66-kDa Zot/zonulin putative receptor is an intracellular modulator of the actin cytoskeleton and of the tight junctional complex, whose activation is triggered by the ligand engagement and subsequent internalization of the ligand/receptor complex. Although our experimental work has been based on either epithelial/endothelial cell monolayers (this paper) or animal whole intestinal tissues [8], there is compelling evidence from Mathan et al. that the distribution of the putative Zot receptor in humans is similar to that described in other mammals [16]. These authors have shown intestinal mucosa ultrastructural changes in adult humans a¡ected by cholera. Jejunal biopsies obtained during the acute phase of the disease showed a marked widening of the lateral intercellular spaces that was present only in the upper third of the villi, and was maximal at the villous tips, gradually decreasing towards the middle of the villus [16]. The availability of puri¢ed receptors will facilitate further studies on the pathophysiological role of the Zot/zonulin system in eukaryotic cells and on the intracellular signaling leading to the modulation of intercellular tj.

[3]

[4]

[5]

[6]

[7]

[8]

[9]

[10]

[11]

[12]

[13]

[14]

[15]

[16]

References

[17]

[1] Claude, P. and Goodenough, D.A. (1973) Fracture faces of zonulae occludentes from `tight' and `leaky' epithelia. J. Cell Biol. 58, 390^ 400. [2] Magnuson, T., Jacobson, J.B. and Stackpole, C.W. (1978) Relation-

5

ship between intercellular permeability and junction organization in the preimplantation mouse embryo. Dev. Biol. 67, 214^224. Madara, J.L. and Pappenheimer, J.R. (1987) Structural basis for physiological regulation of paracellular pathways in intestinal epithelia. J. Membr. Biol. 100, 149^164. Milks, L.C., Conyers, G.P. and Cramer, E.B. (1986) The e¡ect of neutrophil migration on epithelial permeability. J. Cell Biol. 103, 2729^2738. Fasano, A., Baudry, B., Pumplin, D.W., Wasserman, S.S., Tall, B.D., Ketley, J.M. and Kaper, J.B. (1991) Vibrio cholerae produces a second enterotoxin, which a¡ects intestinal tight junctions. Proc. Natl. Acad. Sci. USA 88, 5242^5246. Fasano, A., Fiorentini, C., Donelli, G., Uzzau, S., Kaper, J.B., Margaretten, K., Ding, X., Guandalini, S., Comstock, L. and Goldblum, S.E. (1995) Zonula occludens toxin modulates tight junctions through protein kinase C-dependent actin reorganization, in vitro. J. Clin. Invest. 96, 710^720. Uzzau, S., Cappuccinelli, P. and Fasano, A. (1999) Expression of Vibrio cholerae zonula occludens toxin and analysis of its subcellular localization. Microb. Pathog. 27, 377^385. Fasano, A., Uzzau, S., Fiore, C. and Margaretten, K. (1997) The enterotoxic e¡ect of zonula occludens toxin on rabbit small intestine involves the paracellular pathway. Gastroenterology 112, 839^846. Herrera, R. and Shivers, B.D. (1994) Expression of K1-chimaerin (rac-1 GAP) alters the cytoskeletal and adhesive properties of ¢broblasts. J. Cell Biochem. 56, 582^591. Fasano, A. (1998) Cellular microbiology : how enteric pathogens socialize with their intestinal host. J. Pediatr. Gastroenterol. Nutr. 26, 520^532. Fasano, A., Not, T., Wang, W., Uzzau, S., Berti, I., Tommasini, A. and Goldblum, S.E. (2000) Zonulin, a newly discovered modulator of intestinal permeability, and its expression in Coeliac Disease. Lancet 355, 1518^1519. Fasano, A., Wang, W. and Nie, W. (1998) Isolation and functional characterization of Zonulin, a physiologic modulator of tight junction. Gastroenterology 114, A1141. Marcial, M.A., Carlson, S.L. and Madara, J.L. (1984) Partitioning of paracellular conductance along the ileal crypt^villus axis: a hypothesis based on structural analysis with detailed consideration of tight junction structure^function relationships. J. Membr. Biol. 80, 59^70. Dong, J.M., Smith, P., Hall, C. and Lim, L. (1995) Promoter region of the transcriptional unit for human K1-chimaerin, a neuron-speci¢c GTPase-activating protein for p21rac. Eur. J. Biochem. 227, 636^646. Ridley, A.J., Paterson, H.F., Johnston, C.L., Diekmann, D. and Hall, A. (1992) The small GTP-binding protein rac regulates growth factor-induced membrane ru¥ing. Cell 70, 401^410. Mathan, M.M., Chandy, G. and Mathan, V.I. (1995) Ultrastructural changes in the upper small intestinal mucosa in patients with cholera. Gastroenterology 109, 422^430. Maina, C.V., Riggs, P.D., Grandea III, A.G., Slatko, B.E., Moran, L.S., Tagliamonte, J.A., McReynolds, L.A. and Guan, C.D. (1988) An Escherichia coli vector to express and purify foreign proteins by fusion to and separation from maltose-binding protein. Gene 74, 365^373.

FEMSLE 9697 21-12-00