Clostridium perfringens Enterotoxin Utilizes Two Structurally Related Membrane Proteins as Functional Receptors in Vivo *

Human and mouse cDNAs showing homology to theClostridium perfringens enterotoxin (CPE) receptor gene (CPE-R) from Vero cells (DDBJ/EMBL/GenBankTMaccession no. D88492) (Katahira, J., Inoue, N., Horiguchi, Y., Matsuda, M., and Sugimoto, N. (1997) J. Cell Biol. 136, 1239–1247) were cloned. They were classified into two groups, the Vero cell CPE receptor homologues and rat androgen withdrawal apoptosis protein (RVP1; accession no. M74067) homologues, based on the similarities of primary amino acid sequences. L929 cells that were originally insensitive to CPE became sensitive to CPE on their transfection with cDNAs encoding either the CPE receptor or RVP1 homologues, indicating that these gene products are not only structurally similar but also functionally active as receptors for CPE. By binding assay, the human RVP1 homologue showed differences in affinity and capacity of binding from those of the human CPE receptor. Northern blot analysis showed that mouse homologues of the CPE receptor and RVP1 are expressed abundantly in mouse small intestine. The expression ofCPE-R mRNA in the small intestine was restricted to cryptic enterocytes, indicating that the CPE receptor is expressed in intestinal epithelial cells. These results are consistent with reports that CPE binds to the small intestinal cells via two different kinds of receptors. High levels of expression of CPE-R and/orRVP1 mRNA were also detected in other organs, including the lungs, liver, and kidneys, but only low levels were expressed in heart and skeletal muscles. These results indicate that CPE uses structurally related cellular proteins as functional receptors in vivo and that organs that have not so far been recognized as CPE-sensitive have the potential to be targets of CPE.

The enterotoxin produced by Clostridium perfringens (CPE) 1 is a simple protein with a molecular weight of ϳ35,000. Known as a causative agent of diarrhea, this organism (1) elicits fluid accumulation in the intestinal tract by altering the membrane permeability of intestinal epithelial cells (2,3). Pore formation in the cytoplasmic membrane is now accepted as the underly-ing mechanism of its effect (4 -7).
Not only humans, but also various experimental animals have been shown to be sensitive to CPE (8), suggesting that the sensitivity is not restricted to a particular species. Although the natural target of CPE is the intestine, CPE has also been detected in other tissues and organs, including the liver and kidneys, after its intravenous injection into rats and mice (9). In addition, cultured cells of the intestine, liver, and kidneys from various species have been shown to be sensitive to CPE (10 -14). Two different receptors with high and low affinity to CPE have been found on the surface of rabbit intestinal epithelial cells (10). Horiguchi et al. (14) showed that Vero cells and Madin-Darby canine kidney cells, both of which are derived from kidneys, express high and low affinity receptors, respectively. Since the cytotoxic action of CPE to target cells requires its binding to specific receptors (14,15), at least two molecules with different affinities to CPE are considered to exist in various organs of a wide range of species.
Recently, we cloned a cDNA for the CPE receptor (CPE-R) from a CPE-sensitive Vero cell cDNA library (16). The cDNA encodes a highly hydrophobic transmembrane protein of ϳ22 kDa, the physiological functions of which have not yet been elucidated. The amino acid sequence of the Vero cell CPE receptor showed close similarity to that of the rat androgen withdrawal apoptosis protein RVP1 (17). The Vero cell CPE receptor corresponds to the reported high affinity binding site for CPE. CPE-R was found to be expressed in CPE-sensitive cell lines from different origins; i.e. Vero cells from monkey kidneys, Henle intestine 407 cells from human small intestine, and Hep3B cells from human liver. In contrast to these sensitive cells, CPE-insensitive cells, such as the human erythroleukemia cell line K562 and human lymphoblastoid cell line JY, were found not to express CPE-R, suggesting that the expression of CPE-R is tissue-and/or organ-specific.
In this study, to obtain further insight into the receptors for CPE in vivo, we sought for human and mouse cDNAs similar to the Vero cell CPE-R. Here we report that these gene products can be classified into two groups homologous to the CPE receptor and RVP1, based on their similarities in structure to the Vero cell CPE receptor and RVP1 and differences in their affinities and binding capacity to CPE, and that they both have the ability to confer CPE sensitivity to the insensitive L929 cell line. We also found that the expressions of CPE-R and RVP1 were observed in a wide variety of organs, including the small intestine. These data indicate that two different gene products are expressed in target organs in vivo and are probably recognized as different affinity receptors by CPE.

EXPERIMENTAL PROCEDURES
Identification of Mouse and Human Homologues of CPE-R and RVP1, and Plasmid Construction-Human and mouse genes showing homology to the Vero cell CPE-R were identified by searching the dbEST data base with the BLASTN search program (18) using default setting via the worldwide web. The accession numbers of the 5Ј-and 3Ј-sequences and the tissue origins of the human expressed sequence tag (EST) clones are summarized in Table I. These clones were isolated from various organs, including infant brain, liver/spleen, lungs, breast, heart, uterus, and colon mucosa.
I.M.A.G.E. cDNA clones 25804, 303411, and 214937 (19) were selected based on their homology to Vero cell CPE-R and were purchased from Research Genetics Inc. (Huntsville, AL).
EcoRI-NotI fragments of the clones were treated with T4 DNA polymerase and introduced into pBluescript SK(Ϫ), which had been linearized with EcoRV and treated with alkaline phosphatase. For sequencing, nested deletion mutants of these clones were prepared using a double-stranded, nested deletion kit (Pharmacia Biotech, Uppsala, Sweden).
For construction of expression vectors, EcoRI-NotI fragments of clone 25804 and 303411 (ϳ1.7 and 1.8 kb, respectively) were inserted into the same site of the pMEPyori18sf Ϫ (20) or pMEneo (21) vectors, and the resulting plasmids were designated as pMEhCPE-R or pMEmCPE-R and pMENeohCPE-R, respectively.
Clone 214937 seemed to be truncated at an internal NotI site within the open reading frame (ORF), since the translational termination codon, a putative poly(A) signal, and poly(A) tail were not identified at the 3Ј-end of this clone. The 3Ј-end of clone 214937 was thus cloned by the 3Ј-rapid amplification of cDNA ends (RACE) method (22). Human poly(A ϩ ) RNA from lung (CLONTECH) was reverse transcribed with a RACE primer (5Ј-CGGACTCGAGTCGACATCGATTTTTTTTTTTTT-TTTTT-3Ј). The resulting first strand cDNA was then amplified by the polymerase chain reaction with a human RVP1 primer (5Ј-TCTCGC-CGCCCTGCTCACCC-3Ј; nucleotides 587-606 of clone 214937) and an adaptor primer (5Ј-CGGACTCGAGTCGACATCGAT-3Ј) for 35 cycles in an Expand Long polymerase chain reaction system (Böehringer Mannheim). A DNA fragment of ϳ0.7 kb was introduced into pBluescript SK(Ϫ) that had been linearlized with EcoRV followed by alkaline phosphatase treatment. Nucleotide sequence analysis revealed that the DNA fragment contained a sequence partially overlapping with that of clone 214937 and a NotI site in its 5Ј-portion, and encoded an amino acid sequence that was highly homologous to the C-terminal half of rat RVP1 (see Fig. 1). Intact cDNA (ϳ1.3 kb) was constructed by ligating the clone 214937 and the 3Ј-RACE product at the unique NotI site and was designated as human RVP1 (hRVP1; accession no. AB000714). The deduced amino acid sequence of human RVP1 was shorter than that of rat RVP1, because of the presence of a translational termination codon at nucleotides 859 -861. Several independent clones of the 3Ј-RACE product as well as a mouse EST clone (accession no. W30232), which was highly homologous to rat RVP1, and a mouse RVP1 homologue isolated from mouse liver (see below) also had an in-frame termination codon at this position (data not shown), indicating that the in-frame termination codon was not due to sequencing errors.
A 3Ј-portion of mouse RVP1 was amplified by reverse transcriptonpolymerase chain reaction (23). A total RNA sample from mouse liver was prepared with a Quick Prep total RNA extraction kit (Pharmacia) and reverse transcribed with a random hexamer. The resulting cDNA was amplified with degenerate primers: RVP1 consensus primer 1 (5Ј-TAYTCYGCGCCGCGAYTCCACC-3Ј; Y is mixture of T and C) and RVP1 consensus primer 2 (5Ј-GAAGGGYGAGGTTTCWYWGTCC-3Ј; W is a mixture of A and T), which contain the consensus sequence of human RVP1 (nucleotides 790 -810 and 1082-1103, respectively) and a putative mouse RVP1 homologue (I.M.A.G.E. clone 349696; accession no. W30232). The nucleotide sequence of the amplified fragment (accession no. AB000715) showed higher diversity from mouse CPE-R (51.3% homology) than from rat RVP1 (91.1% homology) and was cloned into pBluescript SK(Ϫ). The resulting plasmid was designated as pBSmRVP3Ј.
For construction of the human RVP1 expression vectors pMEhRVP1 and pMENeohRVP1, the EcoRI-XbaI fragment of hRVP1 was introduced into EcoRI-XbaI digested pMEPyori18sf Ϫ and pMEneo (21).
Cell Culture and Establishment of a Stable Cell Line-Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum at 37°C under 5% CO 2 in air. L929 cells stably expressing human CPE receptor and RVP1 were established essentially as described previously (16). Typical cell lines were designated as NeohCPE-R and NeohRVP1.
Sequence Analysis-The nucleotide sequences of both strands of the cloned genes were determined by the dideoxy chain termination method using Thermo Sequenase (Amersham International plc, United Kingdom). Multiple alignment of the deduced amino acid sequences was performed by the CLUSTAL method (24) and output printing was made by the BOXSHADE program.
Northern Blot Analysis-The expression of CPE-R and RVP1 in various tissues of mice was examined by Northern blot analysis performed on Mouse Multiple Tissue Northern blot (CLONTECH). An EcoRI-SacI fragment containing the entire coding region and a part of the 5Ј-untranslated region of pMEmCPE-R or an EcoRI-HincII fragment of pBSmRVP3Ј containing a part of the coding region and 3Јuntranslated region of mouse RVP1 were radiolabeled with [ 32 P]dCTP and used as probes. Hybridization was performed in 5 ϫ SSPE (1 ϫ SSPE is 180 mM NaCl, 10 mM sodium phosphate buffer (pH 7.7), 1 mM EDTA) containing 10 ϫ Denhardt's solution (1 ϫ Denhardt's solution is 0.02% bovine serum albumin (BSA), 0.02% Ficoll 400, 0.02% polyvinylpyrrolidone), 2% SDS, 50% formamide, and 0.1 mg/ml salmon sperm DNA at 42°C for 12 h. The membrane was washed once with 2 ϫ SSPE, 0.05% SDS at room temperature and twice with 0.1 ϫ SSPE, 0.1% SDS at 50°C, and then exposed to an imaging plate for 20 h. Radioactive bands were visualized with a Bioimaging analyzer (Fuji Film).
Total RNA was isolated from mouse intestine and L929 cells as described above. Each sample was size-fractionated on 6.3% formaldehyde, 1.2% agarose gel and transferred to a charged nylon membrane (Hybond Nϩ, Amersham). Hybridization and washing were done as described above. The blots were rehybridized with 32 P-labeled human EF-1␣ cDNA probe (25) to confirm the amounts and integrities of the samples.
Flow Cytometric Analysis, 125 I-CPE Binding Assay, and Cytotoxicity Assay-Flow cytometric analysis with a biotinylated CPE C-terminal fragment (H 10 PER) was performed as described previously (16). Briefly, L929pyT18 cells were transfected with 10 g of either pMEPyori18sf Ϫ , pMEhCPE-R, pMEmCPE-R, or pMEhRVP1. After 48 h, the cells were harvested and resuspended in 100 l of phosphate-buffered saline (PBS) containing 1% BSA (PBS-BSA). Then the cells were treated with biotinylated H 10 PER (10 g/ml), followed by phycoerythrin-conjugated streptavidin (20 g/ml, Biomeda Corp., Foster City, CA). They were then washed with PBS-BSA, and their fluorescence was examined in a FACScan (Bekton Dickinson, Mountain View, CA).
For testing the CPE sensitivity of transfected cells, purified CPE (26) was added to the culture medium at a concentration of 100 ng/ml. Radioiodination of CPE, binding assay, and Scatchard analysis were performed as described previously (16).

Identification and Sequence Analysis of cDNAs Showing Homology to Vero Cell CPE-R-I.M.A.G.E. cDNA clones 25804
from an infant human brain library, which showed the highest homology to Vero cell CPE-R and clone 303411 from a fetal mouse library were sequenced, and the deduced amino acid sequences were compared with that of the Vero cell CPE receptor. They contained small ORFs encoding 209 (clone 25804) and 210 (clone 303411) amino acids, respectively. The amino acid sequences of these clones were highly homologous (clone 25804, 99.0% homology; clone 303411, 83.8% homology) to that of the Vero cell CPE receptor (Fig. 1A, vCPE-R) and had four highly hydrophobic domains with potentials to form transmembrane domains (Fig. 1B). Thus we designated these genes as hCPE-R (clone 25804; accession no. AB000712) and mCPE-R (clone 303411; accession no. AB000713). The ORF of clone 214937 was truncated, so the complete ORF was deduced from the nucleotide sequences of the clone 214937 and the 3Ј-RACE product (see "Experimental Procedures"). It consisted of 220 amino acids and also contained four putative transmembrane domains (Fig. 1B). The amino acid sequence of this ORF showed higher homology to that of the rat RVP1 (89.4% identity, 97.7% similarity) than to that of the human CPE receptor (69.9% identity, 96.7% similarity) or the Vero cell CPE receptor (70.9% identity, 96.6% similarity). We, therefore, named this gene encoding the human homologue of RVP1 as hRVP1. The amino acid sequence of the mouse CPE-R gene product showed higher homology to that of the Vero cell CPE receptor than to that of the rat RVP1 (66.8% identity, 94.6% similarity) or human RVP1 (66.0% identity, 94.3% similarity). Taken together, we concluded that the EST clones which showed similarity to the Vero cell CPE-R could be classified into RVP1 and CPE-R homologues and that these two molecules are not expressed from the same locus by alternative splicing but are encoded on different loci.
The Homologous Genes Encoding the Functional CPE Receptor-We tested whether these gene products actually act as functional receptors for CPE. The human and mouse CPE receptors and the human RVP1 were each transiently expressed in L929pyT18 cells, and flow cytometric analysis using the biotinylated H 10 PER probe was performed as described previously (16). A peak showing increased fluorescent intensity was observed when the human and mouse CPE receptor homologues (Fig. 2, B and C, thick lines) as well as the Vero cell CPE-R cDNA ( Fig. 2A, thick line) were expressed. When human RVP1 was expressed in L929pyT18 cells, cells brighter than the background appeared, but they did not form any separated peaks due to their relatively low fluorescent intensity (Fig. 2D, thick line). This tendency, confirmed by three independent experiments (data not shown), could be due to both its lower affinity and fewer expression on these cells as shown below. When the pMEPyori18sf Ϫ vector was introduced into the cells, no increase of the fluorescent intensity was observed (Fig. 2, A-D, thin lines).
We then tested the CPE sensitivities of L929pyT18 cells transiently expressing the CPE receptor homologues and hu-  man RVP1. Balloon formation and cell lysis were observed in cells expressing human (Fig. 3A) or mouse (Fig. 3B) CPE receptors or human RVP1 (Fig. 3C) after treatment with purified CPE (note that only ϳ30% of the cells exhibited CPE sensitivity, since the transfection efficiency reached 30% at maximum), while cells transfected with the vector alone did not show the typical morphological abnormalities induced by CPE (Fig. 3D). Thus we concluded that both human and mouse CPE receptors, and human RVP1 are functional as receptors for CPE.
Different Kinetics of Binding of CPE to Human RVP1-Previously we reported that CPE binds to the Vero cell CPE receptor with high affinity (apparent affinity constant, 1.49 ϫ 10 8 M Ϫ1 ) (16). To compare the binding kinetics of CPE to the CPE receptor with that to RVP1, both of which are derived from the same species, we established the NeohCPE-R and NeohRVP1 cell lines stably expressing the human CPE receptor and RVP1 and performed binding assay using 125 I-labeled CPE. Equilibrium binding data (Fig. 4, A and B, insets) analyzed by Scatchard plots (Fig. 4, A and B) showed the presence of a single order of binding sites on each cell line with an apparent affinity constant value of 7.94 ϫ 10 7 M Ϫ1 for human CPE receptor and 4.60 ϫ 10 7 M Ϫ1 for human RVP1. The number of human RVP1 expressed on the cell line was estimated about 8 times fewer than that of the human CPE receptor, even though they were expressed from the same vector.
Expressions of the Mouse CPE-R and RVP1 Gene Products in Various Tissues-Northern blotting revealed the expression of a transcript of CPE-R of about 1.8 kb in total RNA from intestine, but not from the CPE-insensitive L929 cell line (Fig. 5A,  upper panel). The expression of CPE-R in other mouse tissues was also tested. Two kinds of transcripts of different sizes were expressed (Fig. 5B, upper panel). Kidneys expressed a transcript of about 1.8 kb abundantly. This 1.8-kb transcript was also detected in heart and skeletal muscle, although at lower levels. In liver, a shorter transcript of about 1.3 kb was expressed in place of the 1.8-kb transcript. In lungs, both the 1.8and 1.3-kb transcripts were detected. The presence of transcripts of different sizes is consistent with the results of a data base search except for the pattern in mouse liver, in which the 1.3-kb transcript was the major one found. The cDNA fragment corresponding to a part of the 3Ј-noncoding region of mouse RVP1 hybridized with the 1.3-kb transcript, but not with the 1.8-kb transcript in stringent washing conditions, indicating that RVP1 is the major transcript in place of CPE-R in mouse liver (Fig. 5B, middle panel). The expression of RVP1 was also detected in intestine but not in L929 cells (Fig. 5A, middle  panel). The lack of the expression of both CPE-R and RVP1 in L929 cells coincides well with their lack of the sensitivity to CPE. Brain, spleen, heart, skeletal muscle, and kidneys (Fig.  5B, middle panel) also expressed RVP1, but at much lower levels than in intestine, liver, and lungs. No expression of 1.8-kb CPE-R mRNA was detected in brain or spleen.

CPE-R mRNA in Small Bowel Crypts of an Adult
Mouse-For determination of the histological localization of CPE-R mRNA in mouse small intestine, in situ hybridization was performed. In the jejunum of adult mice, intense positive staining was observed in the cryptic cells when biotin-labeled antisense-CPE-R cRNA probe was used (Fig. 6A). Enterocytes in the lower part of the villi exhibited either faint or, in most cases, no positive signals for CPE-R mRNA. The upper half of the villi showed no positive signals. Cells from the upper to lower parts of the villi and cryptic cells showed no positive signal with biotin-labeled sense-CPE-R cRNA as a probe (Fig.  6B). The probe used in this experiment contained the coding region of the CPE-R gene, and thus could hybridize with both CPE-R and RVP1 mRNAs (see Fig. 5, A and B, upper panels). However, because it hybridized much more stably with CPE-R, the majority of positive signals indicated the presence of CPE-R mRNA.

DISCUSSION
Two Different Kinds of Molecules Act as Receptors for CPE-In the present study we found that the cellular proteins that function as receptor for CPE in vivo could be classified into two different groups of the CPE receptor and RVP1 types. These receptors were very similar in structure but had different affinities to CPE. Also we found that the cell lines expressing human RVP1 showed lower binding capacity to CPE than those expressing human CPE receptor. McDonel (10) and Horiguchi et al. (14) examined different cells (i.e. rabbit intestinal cells and kidney cell lines from monkey and dog) and found receptors of two kinds differing in their affinities to CPE. The fact that the CPE receptor and RVP1 are expressed as two different receptors in intestine and kidneys is consistent with their observations. Some findings, however, require further investigation. In our results, Vero, mouse, and human CPE receptors tended to show higher affinities for CPE than human for RVP1. This implies that the CPE receptor and RVP1 correspond to the high and low affinity receptors in the literature, respectively. The reported differences in the affinities between the high and low affinity receptors are greater than those of our data. This could be attributed to the differences in primary amino acid sequences among the receptors from different species, since even a few amino acid substitution alters the affinity to CPE (see also below). Although further comparative studies of CPE receptors and RVP1 from various species are necessary to well substantiate this hypothesis, our results provide, for the first time, a molecular basis for two kinds of receptors for CPE with different affinities.
If the hydrophilic regions of these receptor molecules are the extracellular domains, either or all of them probably provide a binding site for CPE. The single or multiple amino acid replacement in these regions of the CPE receptors and RVP1 might cause their different affinities to CPE. However, it should be noted that the amino acids in the putative transmembrane segments may also contribute to the interaction with CPE, since the amino acid replacements which occur within the hydrophobic domains (e.g. Thr 9 and Leu 131 of vCPE-R are replaced by Met in hCPE-R; see Fig. 1A) affected their affinities to CPE.
Expressions of the CPE Receptor and RVP1-In rabbit intestine, the brush border membrane is reported to have affinity and sensitivity to CPE (27). Thus the CPE receptor and RVP1 must be expressed in mature intestinal epithelial cells on the tip of the villi. However, we found that only the cryptic enterocytes express detectable levels of the CPE-R mRNA. Interestingly, the sensitivities of Vero and HeLa cells to CPE were reported to increase on treatment of the cells with interferon (28) or serum starvation. 2 In addition we found that CPE-R is expressed in infant human brain, but not detectable in adult mouse brain. These observations prompt us to consider that the expression of CPE-R may be regulated during the course of cell growth and/or differentiation. Thus the notion that the expression of CPE-R is regulated during the differentiation and/or proliferation process of enterocytes along the crypt-villus axis possibly explains the expression pattern of the CPE-R transcript in intestine.
McDonel and McClane (15) reported that high and low affinity receptors are both expressed by a single cell line (Vero cells). However, we previously observed only a 1.8-kb CPE-R transcript in these cells (16). Possibly Vero cells express RVP1 as a low affinity receptor, but its hybridization to the CPE-R probe was too unstable to detect its expression (note that the signal intensity of RVP1 transcript obtained with the CPE-R probe was much weaker than that obtained with the RVP1 probe; see Fig. 5B). Northern blot analysis of the expression of RVP1 mRNA in Vero cells using RVP1 probe will solve this problem. Further examination of whether CPE-R and RVP1 are expressed concomitantly in the same cell population or separately in specific cell types will also provide a clue for understanding the physiological functions of these two proteins.
We detected the expressions of CPE-R and RVP1 in liver. Previously we demonstrated that the CPE-R transcript was expressed in CPE-sensitive Hep3B cells, which are derived from human liver (16). Mouse liver cells in primary culture were CPE-sensitive but exhibited lower affinity to CPE than did Vero cells (11). Since we showed that mouse liver expresses the RVP1 transcript abundantly, it is probably recognized as the target molecule for CPE. The expression of CPE-R in human liver was confirmed by the data base search shown in Table I and Northern blot analysis of Hep3B cells, so the possibility that some portion of human hepatocytes may express RVP1 under particular conditions cannot be excluded.
We observed the expression of CPE-R in lungs, heart, and skeletal muscles as well as in intestine, liver, and kidneys. The former organs have not so far been thought to be sensitive to CPE in vivo. Cell lines derived from rat skeletal muscle (L-6 cells) and rat smooth muscle (A-10 cells) were CPE-insensitive. 3 Moreover, we found that a mouse myoblastic cell line C2C12 did not express CPE-R mRNA or showed any morphological alterations by the addition of CPE to the culture medium; even when these cells were induced to differentiate into multinucleate myotubes, they did not show CPE sensitivity (data not shown). We think that the reported insensitivity of the cultured myogenic cell lines may indicate that they do not express sufficient amounts of CPE-R or RVP1 to be sensitive to CPE. It is possible that some accessory molecules that are essential for CPE-induced cytotoxicity are not present in myogenic cells, resulting in their insensitivity. However, this possibility is unlikely, since a C2C12 cell line stably expressing the Vero cell CPE-R became sensitive to CPE (data not shown). The mechanisms regulating expressions of the CPE receptor and RVP1 in specific cells in vivo remain to be elucidated. Breast R55031 a Tissue origin of each library.