Distinct roles for the N- and C-terminal regions in the cytotoxicity of pierisin-1, a putative ADP-ribosylating toxin from cabbage butterfly, against mammalian cells

Abstract

Pierisin-1 is an 850-aa cytotoxic protein found in the cabbage butterfly, Pieris rapae, and has been suggested to consist of an N-terminal region with ADP-ribosyltransferase domain and of a C-terminal region that might have a receptor-binding domain. To elucidate the role of each region, we investigated the functions of various fragments of pierisin-1. In vitro expressed polypeptide consisting of amino acid residues 1–233 or 234–850 of pierisin-1 alone did not show cytotoxicity against human cervical carcinoma HeLa cells. However, the presence of both polypeptides in the culture medium showed some of the original cytotoxic activity. Introduction of the N-terminal polypeptide alone by electroporation also induced cell death in HeLa cells, and even in the mouse melanoma MEB4 cells insensitive to pierisin-1. Thus, the N-terminal region has a principal role in the cytotoxicity of pierisin-1 inside mammalian cells. Analyses of incorporated pierisin-1 indicated that the entire protein, regardless of whether it consisted of a single polypeptide or two separate N- and C-terminal polypeptides, was incorporated into HeLa cells. However, neither of the terminal polypeptides was incorporated when each polypeptide was present separately. These findings indicate that the C-terminal region is important for the incorporation of pierisin-1. Moreover, presence of receptor for pierisin-1 in the lipid fraction of cell membrane was suggested. The cytotoxic effects of pierisin-1 were enhanced by previous treatment with trypsin, producing “nicked” pierisin-1. Generation of the N-terminal fragment in HeLa cells was detected after application of intact entire molecule of pierisin-1. From the above observations, it is suggested that after incorporation of pierisin-1 into the cell by interaction of its C-terminal region with the receptor in the cell membrane, the entire protein is cleaved into the N- and C-terminal fragments with intracellular protease, and the N-terminal fragment then exhibits cytotoxicity.

We recently found that a potent cytotoxic factor against human cancer cells is present in the body fluids of larvae and pupae of cabbage butterfly, Pieris rapae (1). The factor responsible for this cytotoxicity has been identified as a 98-kDa protein, and we named it pierisin-1 (2). Subsequent studies showed that pierisin-1 exhibited potent cytotoxic effects against various types of human cancer cell lines and human umbilical vein endothelial cells, with IC₅₀ values ranging from 0.043 ng/ml to 150 ng/ml, and induced typical apoptosis of most cell lines up to 48 h (3). The pierisin-1 gene has been cloned from P. rapae larval mRNA. Pierisin-1 shows regional sequence similarities with ADP-ribosylating toxins such as the A-subunit of cholera toxin. Disruption of a possible NAD-binding site by site-directed mutagenesis results in the loss of its cytotoxic activity (4). The mRNA of pierisin-1 is highly expressed in fifth-instar larvae, and the concentration of pierisin-1 reached highest before pupation (4). From these observations, it was suggested that pierisin-1 might play some important role in the pupation of P. rapae through the induction of apoptosis in some cells in the larvae and pupae. It was also suggested that pierisin-1 would be present as a protective agent against invading organisms such as parasitic Hymenoptera. Another cabbage butterfly, Pieris brassicae, also contains a molecular analog of pierisin-1, and it was named pierisin-2 (5, 6). Cloning of the pierisin-2 gene indicated that pierisin-2 was 91% identical to pierisin-1 at the amino acid level (6).

Trypsin cleaves pierisin-1 specifically at Arg-233–Ser-234, and other proteases also cleave near this site (4). Trypsin-cleaved pierisin-1 has been thought to be “nicked” pierisin-1, which is composed of properly associated N- and C-terminal fragments with a similar structure to that of intact pierisin-1 (4, 7). These observations suggest that N-terminal (27 kDa) and C-terminal (71 kDa) regions of pierisin-1 should have different functions. The N-terminal region shows partial sequence similarity with ADP-ribosylating enzymes, indicating that it could be a catalytic domain. The C-terminal region shares sequence similarity with HA-33 (or HA1), a subcomponent of hemagglutinin of botulinum toxin (8, 9). Recent reports on requirement of sialic acid or galactose for binding ability of HA-33 (10, 11) suggest the existence of a membrane-binding domain in the C-terminal region of pierisin-1.

Understanding of the biological nature of pierisin-1 could provide informative data for elucidation of the significance of pierisin-1 in the cabbage butterfly and of mechanisms of apoptosis in human cancer cells mediated by pierisin-1. Thus, the present study was designed to determine the functions of various fragments of pierisin-1. Our results suggest that after incorporation of pierisin-1 into the cell by interaction of its C-terminal region with the cell membrane, the protein is cleaved by proteolysis into the N- and C-terminal fragments, and the N-terminal fragment then exhibits cytotoxicity. It is also discussed that the presence of membrane receptors for pierisin-1 possibly determines the sensitivity of mammalian cells to this toxin.

Materials and Methods

Cell Lines.

Human cervical carcinoma HeLa cells and mouse melanoma MEB4 cells were obtained from the RIKEN cell bank (Tsukuba, Japan). HeLa cells were the most sensitive, and MEB4 cells were the most resistant to the effect of pierisin-1 so far tested. The IC₅₀ value of the MEB4 was 270 ng/ml, which is around 5,000-fold higher than HeLa cells. Human gastric carcinoma TMK-1 cells were obtained from Dr. E. Tahara (Hiroshima University, Hiroshima, Japan). These three cell lines were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated FBS (GIBCO/BRL). The cell cultures were maintained at 37°C in an atmosphere of 95% air/5% CO₂.

Analysis of Cytotoxic and Apoptosis-Inducing Activities.

The cytotoxic effects of specimens on cultured cells were examined by using the WST-1 [2(4-iodophenyl)-3(4-nitrophenyl)-5(2,4-disulfophenyl)-2H-tetrazolium; Dojindo Laboratories, Kumamoto, Japan] cell proliferation assay as described previously (3). Apoptosis-inducing activity of the specimens was confirmed by morphological analysis of nuclei of cells stained with Hoechst 33342 under a fluorescence microscope.

Visualization of Pierisin-1 by Fluorescence Labeling.

One milligram of purified pierisin-1 from the pupae of P. rapae (2) was subjected to fluorescence labeling with a FluoroLink-Ab Cy2 labeling kit (Amersham Pharmacia) by using the method provided by the manufacturer. Cy2-labeled pierisin-1 was analyzed with regard to its protein concentration, cytotoxicity, and apoptosis-inducing activity. Binding and internalization of the labeled pierisin-1 to cultured cells were observed under a fluorescence microscope.

Preparation of Polypeptides of Pierisin-1 by in Vitro Expression.

A variety of fragments of pierisin-1 were expressed in vitro from an intact pierisin-1 cDNA subclone (4) by using MEGAscript and rabbit reticulocyte lysate (Ambion, Austin, TX) after amplification of the desired coding region by PCR, as described previously (4). To prepare polypeptides that started at amino acid residues 1, 164, 234, 242, and 270, a 5′ primer containing a T7 promoter sequence was used: 5′-TAATACGACTCACTATAGGGCGAATTGCCACCATGGCTGACCGTCAACCTTAC, 5′-TAATACGACTCACTATAGGGCGAATTGCCACCATGGAGGTTGCCTTTCCTGG, 5′-TAATACGACTCACTATAGGGCGAATTGCCACCATGTCAGCCAGCTCTTATGATGACT, 5′-TAATACGACTCACTATAGGGCGAATTGCCACCATGTACGGTGGTACAGGTAATGT, and 5′-TAATACGACTCACTATAGGGCGAATTGCCACCATGATTGAAAGTATAAAAGACAAAAA, respectively. Because amino acid 234 was not methionine, the primers for the polypeptides beginning from position 234 additionally included a methionine codon at a position before 234. Similarly, 3′ primers of 5′-TGGGATTTATCCAAATTCTTT, 5′-ATCTCTCAGAACGTTGATCTCTA, 5′-CGTACATCAAGTCATCATAAGAG, 5′-TCATAAATTCACCAGCTGCAAT, and 5′-GCCCTGTTTCACAATGTATG were used to prepare polypeptides that ended at amino acid residues 195, 233, 243, 270, and 850, respectively. To prepare a polypeptide containing a mutation at Glu-165, mutated E165Q clone (4) was used. The expression efficiency of each protein was estimated by SDS/PAGE of labeled translates. Each translated protein in lysate was added to a culture medium containing HeLa cells, incubated for 72 h, and subjected to a WST-1 cell proliferation assay.

Electroporation.

Exponentially growing cells were harvested by trypsinization and washed once with serum-free Opti-MEM (GIBCO/BRL). The cells were subsequently resuspended at a density of 1 × 10⁷ cells/ml in the same medium, and 40 μl of suspension were transferred to a 0.1-cm Gene Pulser cuvette (Bio-Rad) on ice. A sample of reticulocyte lysate including the translated protein solution was added to the suspension in the cuvette and mixed well. The mixture was then exposed to a single electric pulse of 1,000 V/cm, with a capacitance of 250 μF and a pulse duration of 10–20 ms by using a Bio-Rad gene pulser system. The cuvette was immediately placed on ice for 10 min, and cells were then suspended in RPMI-1640 containing FBS and plated onto wells of 6- and 96-well dishes. Under these conditions, the incorporation of external molecules was confirmed by adding β-galactosidase and staining with X-gal in the culture dishes. One to three days later, the cells were examined for morphological changes under phase-contrast and fluorescence microscopy and subjected to a WST-1 cell proliferation assay.

Microinjection.

Reticulocyte lysate containing the N-terminal polypeptide or control reticulocyte lysate was diluted to 6% with PBS, pH 6.9. These samples were microinjected into HeLa cells, grown on 35-mm Petri dishes, by using a semiautomatic micromanipulator/microinjector (model 5171/5246, Eppendorf). After injection, the medium was changed, and cells were incubated at 37°C.

Western Blotting.

HeLa cells treated with purified pierisin-1 were collected by a cell scraper and sonicated in PBS. The extract was then separated by SDS/PAGE and blotted on a polyvinylidene difluoride membrane (Millipore) by semidry electrophoretic transfer. Rabbit antisera against SDS-denatured pierisin-1 purified from pupae of P. rapae and keyhole limpet hemocyanin-conjugated peptide (CYGFAKNNHPSIFVSTTKTORNKK), corresponding to amino acid residues 101–123 of pierisin-1, were prepared by Sawady Technology (Tokyo). Horseradish peroxidase-conjugated anti-rabbit IgG (Amersham Pharmacia) was used as the secondary antibody. Protein–antibody complexes were visualized by an ECL chemiluminescence kit (Amersham Pharmacia).

Detection of Radioisotope-Labeled Pierisin-1 Fragment.

Cells were treated with the reticulocyte lysate containing in vitro expressed 1–233 N-terminal and/or 234–850 C-terminal polypeptides labeled with [³⁵S]methionine. The cells were then lysed with electrophoresis sample buffer containing SDS and vigorously mixed. The protein samples were separated by SDS/PAGE and visualized by autoradiography.

Cleavage of Pierisin-1 at Arg-233–Ser-234 by Trypsin.

Ten micrograms of purified pierisin-1 were mixed with 1 μg of trypsin (GIBCO/BRL) in 10 μl of 50 mM Tris-HCl (pH 7.5) and 1 mM EDTA, and incubated for 1 h at 37°C. Cleavage was confirmed by SDS/PAGE analysis.

Analysis of Inhibitory Effects of Membrane and Lipid Fractions on the Cytotoxicity of Pierisin-1.

The cultured 2 × 10⁸ of HeLa and MEB4 cells were collected by centrifugation after treatment with buffer containing PBS, 7 mM EDTA, and 0.01% sodium azide at 37°C for 30 min; the membrane fraction was then isolated as previously described (12). Finally, the membrane fraction was suspended in 1 ml of RPMI-1640 medium. The total lipids were extracted from 2 × 10⁸ of harvested HeLa or MEB4 cells according to the method of Murayama et al. (13) and dissolved in 1 ml of RPMI-1640 medium containing 10% dimethyl sulfoxide.

With membrane or lipid fractions thus obtained, their inhibitory effects on the cytotoxicity of pierisin-1 to HeLa cells were examined. A sample of 0.5 ng of native pierisin-1 was preincubated in a 50-μl aliquot solution containing a membrane fraction or total lipid fraction for 1 h at 4°C. The appropriate volume of preincubated mixture was combined with the cultured HeLa cells (1 × 10⁴/well of 96-well plate) and incubated for 15 min at 37°C. Then, cells were washed three times and further incubated for 72 h. Viable cell numbers were assessed by WST-1 cell proliferation assay.

Results

Direct Observation of Binding and Internalization of Fluorescence-Labeled Pierisin-1.

To examine whether pierisin-1 could bind to the membrane and be internalized into cells, the protein was visualized by conjugation with Cy2 fluorescence dye. Labeled Cy2-pierisin-1 exhibited strong cytotoxicity against both TMK-1 and HeLa cells, similar to unlabeled pierisin-1. In addition, both Cy2-pierisin-1 and the unlabeled one indistinguishably induced chromatin condensation and nuclear fragmentation, indicating apoptotic cell death in these cells. Thus, the biological activity of pierisin-1 was considered to be only minimally reduced by fluorescence labeling.

Typical findings on fluorescence microscopy of cancer cells treated with 2 μg/ml Cy2-pierisin-1 are shown in Fig. 1. The high concentration of fluorescence-labeled pierisin-1 allowed the direct observation of incorporation of the protein into cells by microscopic analysis. Large numbers of fluorescent vesicles were detected inside HeLa cells, which are sensitive to pierisin-1, by treatment for 1 h at 37°C. The vesicles tended to accumulate in the perinuclear region (Fig. 1A). When treatment was performed at 4°C, most of the fluorescence remained on the cell surface (Fig. 1B). By further incubation of the cells at 37°C, in which Cy2-pierisin-1 bound on the cell surface, fluorescent vesicles transferred inside the cells (Fig. 1C). In contrast, only weak fluorescence emission was observed both inside and on the surface of Cy2-pierisin-1-treated MEB4 cells, which are resistant to the effect of pierisin-1 (Fig. 1D). These results indicate that pierisin-1 binds to and is readily internalized into HeLa but not MEB4 cells.

Figure 1.

Fluorescence micrographs of Cy2-pierisin-1-treated HeLa and MEB4 cells. (A) HeLa cells were incubated with 2 μg/ml Cy2-pierisin-1 for 1 h at 37°C. (B) HeLa cells were incubated with the same concentration of Cy2-pierisin-1 for 1 h at 4°C. (C) The medium containing Cy2-pierisin-1 was then removed, and the cells were further incubated in a medium free of Cy2-pierisin-1 for 1 h at 37°C. (D) MEB4 cells were incubated with 2 μg/ml Cy2-pierisin-1 for 1 h at 37°C.

Cooperation of the N- and C-Terminal Polypeptides in the Cytotoxicity of Pierisin-1.

It has been suggested that pierisin-1 consists of a Met-1–Arg-233 N-terminal region, which harbors an ADP-ribosyltransferase domain, and a Ser-234–Met-850 C-terminal region, which contains a membrane-binding domain (4). Thus, the functions of various polypeptides of pierisin-1 were examined. As shown in Fig. 2, neither the 1–233 N-terminal polypeptide nor the 234–850 C-terminal polypeptide, which each was expressed in vitro in lysates of rabbit reticulocytes, showed cytotoxic effects in HeLa cells. However, when both polypeptides were present together, cytotoxicity was restored. This cytotoxic effect was about 10% of that of full-length pierisin-1 (see Fig. 2). Cell death induced by a mixture of these polypeptides was confirmed to be apoptotic cell death, as in the case of the in vitro expressed full-length pierisin-1. In contrast, no other two-polypeptide combination (1–195 plus 164–850, 1–243 plus 242–850, or 1–270 plus 270–850) showed cytotoxic activity. The cytotoxicity of the N-terminal polypeptide in the presence of the C-terminal polypeptide was not observed when an E165Q mutant clone (4) was used to express the N-terminal polypeptide.

Figure 2.

Cytotoxicity of the N- and C-terminal polypeptides of pierisin-1 against HeLa cells. Each translated protein in lysate was added to a culture medium containing HeLa cells and incubated for 72 h. Horizontal bars represent polypeptides of pierisin-1. Dashed vertical line is the trypsin cleavage site. The 165th glutamic acid residue is considered to be the putative NAD binding site and indicated as “E.” The mutation site by replacement of the 165th glutamic acid with glutamine is indicated as “Q” (4). (I) The reticulocyte lysate containing 1–850 full-length pierisin-1; the concentration of lysate in culture medium was 0.04%. (II) 1–233 N-terminal polypeptide, >6.4%. (III) 234–850 C-terminal polypeptide, >6.4%. (IV) 1–233 N-terminal plus 234–850 C-terminal polypeptides, 0.02% N + 0.06% C. (V) E165Q 1–233 N-terminal polypeptide, >6.4%. Under the above conditions, about 50% of cell death was detected. The translation efficacy of each polypeptide in rabbit reticulocyte lysate used in this study generally depended on its length. The amounts of the 1–233 N-terminal and 234–850 C-terminal polypeptides in lysate were around 10- and 3-fold greater than that of 1–850 full-length pierisin-1, respectively. Consequently, the activity of the mixture of N- plus C-terminal polypeptides was estimated to be about one-tenth of that of the full-length protein.

Principal Role of the N-Terminal Region in Cytotoxicity in Cultured Cells.

Although the N-terminal region has been suggested to harbor an ADP-ribosyltransferase domain, it alone showed no cytotoxic activity in HeLa cells. Therefore, experiments were designed to incorporate the N-terminal polypeptide into the cells by electroporation, and its cytotoxic activity was investigated. As shown in Fig. 3A, cytotoxicity assays demonstrated that the N-terminal polypeptide, but not the C-terminal polypeptide, exhibited dose-dependent cytotoxicity in HeLa cells. However, no cytotoxicity was observed when the mutated E165Q N-terminal polypeptide was used. Moreover, it should be emphasized that the electroporated N-terminal polypeptide efficiently induced cell death of pierisin-1-resistant MEB4 cells, with almost the same potency as in pierisin-1-sensitive HeLa cells (Fig. 3B). Fig. 4 shows morphological changes in HeLa and MEB4 cells that incorporated the N- and C-terminal polypeptides of pierisin-1 by electroporation. Both cells that were electroporated with reticulocyte lysate containing the 1–233 N-terminal polypeptide apparently detached from the dishes, indicating cell death (Fig. 4 B and F). Fluorescence microscopic analysis of these cells indicated that cell death was apoptotic in nature (Fig. 4E). When HeLa cells were electroporated with control reticulocyte lysate without any translated proteins or lysate containing the 234–850 C-terminal polypeptide, little or no morphological deterioration was observed (Fig. 4 A and C). Under the conditions of electroporation adopted in this experiment, damage to cells was minimal (Fig. 4D). In addition, microinjection of the 6% lysate containing the N-terminal polypeptide into HeLa cells also led to cell death, as in the case of electroporation, but that of control lysate did not. Thus, we conclude that the N-terminal region itself was toxic in HeLa and MEB4 cells, after incorporation into cells.

Figure 3.

Dose-dependent cytotoxic effects of the N- and C-terminal polypeptides of pierisin-1. Cells were electroporated with various concentrations of lysates, incubated for 72 h, and subjected to a cell proliferation assay. ■, control lysate without any translated proteins; ●, lysate containing 1–233 N-terminal polypeptide; □, 234–850 C-terminal polypeptide; ○, N-terminal E165Q polypeptide. (A) HeLa cells. (B) MEB4 cells. Each experiment was carried out twice, and an average value of proliferation activity at each concentration of lysate in electroporation cuvette is given.

Figure 4.

Morphological changes in HeLa and MEB4 cells that incorporated the N- and C-terminal polypeptides of pierisin-1 by electroporation. HeLa cells were electroporated with 1.6% control lysate alone (A), with 1.6% lysate containing 1–233 N-terminal polypeptide (B and E), with 1.6% lysate containing 234–850 C-terminal polypeptide (C), or without lysate (D). (F) MEB4 cells were electroporated with 1.6% lysate containing 1–233 N-terminal polypeptide. Electroporated cells were incubated for 18 h at 37°C. (A–D, F) Phase-contrast micrographs. (E) Fluorescence micrograph of Hoechst 33342-stained cells.

Requirement of the C-Terminal Region of Pierisin-1 for Incorporation.

Fluorescence-labeled pierisin-1 was efficiently incorporated into HeLa cells. To clarify which part of pierisin-1 was required for this incorporation, Western blot analysis of HeLa cells incubated with native pierisin-1 was performed. An immunoreactive band corresponding to full-length pierisin-1 was observed after 15 min of incubation (Fig. 5A). This result indicates that the entire pierisin-1 was incorporated into HeLa cells. Several immunoreactive bands smaller than full-length pierisin-1 also appeared upon the further incubation of HeLa cells without pierisin-1. Among them, a 27-kDa band was confirmed to be the N-terminal fragment by using an N-terminal region-specific antibody (Fig. 5A). Therefore, generation of the N-terminal fragment from intact pierisin-1 in HeLa cells was also evident. Next, HeLa cells were incubated with the lysate containing [³⁵S]methionine-labeled N- and C-terminal polypeptides, and this was followed by an analysis of incorporation (Fig. 5B). Both the N- and C-terminal polypeptides were incorporated when they coexisted. However, neither of the fragments was detected when each was present alone. Thus, both the N- and C-terminal regions are required for the incorporation of pierisin-1 into HeLa cells.

Figure 5.

Detection of various incorporated pierisin-1 polypeptides. (A) Western blot analysis of HeLa cells treated with native pierisin-1. Cells were incubated with 1 μg/ml pierisin-1 for 0 min (lane 1) and 15 min (lane 2) at 37°C. Medium containing pierisin-1 was then removed from the culture, and the cells were further incubated without pierisin-1 for another 45 min (lane 3) and 3 h (lane 4). Anti-purified pierisin-1 (full-length) and anti-peptide corresponding to amino acids 101–123 of pierisin-1 (N-specific) antibodies were used to detect pierisin-1. (B) SDS/PAGE of radioisotope-labeled pierisin-1 fragment incorporated into HeLa cells. Cells were incubated with 2% lysate containing the [³⁵S]methionine-labeled C-terminal polypeptide plus 2% control lysate (*C) or plus 2% lysate containing the nonradioisotopic N-terminal polypeptide (*C + N) for 0, 30, and 180 min. Cells were also incubated with 2% lysate containing [³⁵S]methionine-labeled 1–233 N-terminal polypeptide plus 2% control lysate (*N) or plus 2% lysate containing nonradioisotopic 234–850 C-terminal polypeptide (*N + C) for 0, 30, and 180 min.

Enhancement of the Cytotoxic Activity for Pierisin-1 by Cleavage with Trypsin.

Next, native pierisin-1 was treated with trypsin, and its cytotoxicity against HeLa cells was assessed. Trypsin-cleaved “nicked” pierisin-1 showed increased cytotoxicity against the cells. The IC₅₀ of trypsin-cleaved pierisin-1 against HeLa cells was 0.015 ng/ml, whereas that of intact pierisin-1 was 0.06 ng/ml. Therefore, the activation of pierisin-1 by cleavage between the N- and C-terminal regions was demonstrated.

Existence of the Possible Receptor of Pierisin-1 in the Membrane Fraction of HeLa Cells.

Addition of a receptor molecule for pierisin-1 is expected to competitively inhibit cytotoxicity of the protein to the cells by its binding. Therefore, membrane fractions were collected from the HeLa and MEB4 cells, and their inhibitory effects on the cytotoxicity of the pierisin-1 were examined. Treatment of HeLa cells with 2 ng/ml pierisin-1 at a pulse duration of 15 min induced cell death in about 50% of the cells. Contrary to this, preincubation of 2 ng/ml pierisin-1 with membrane fraction from HeLa cells before the treatment caused cell death in only 20% (Fig. 6). On the other hand, addition of membrane fraction from MEB4 cells did not affect the cytotoxicity of pierisin-1. Next, total lipid fractions were prepared from HeLa and MEB4 cells, and their inhibitory effects on cytotoxicity of pierisin-1 to HeLa cells were tested. The cytotoxicity of pierisin-1 was decreased by addition of total lipid fraction from HeLa cells to almost the same extent as in the case of membrane fractions. However, total lipid fraction from MEB4 cells showed almost no effect on the cytotoxicity.

Figure 6.

Effect of membrane fraction from HeLa or MEB4 cells on cytotoxicity of pierisin-1 to HeLa cells. A sample of 0.5 ng of native pierisin-1 was preincubated in a 50-μl aliquot solution containing membrane fraction from HeLa and MEB4 cells for 1 h at 4°C. The appropriate volume of preincubated mixture was combined with the cultured medium of HeLa cells and incubated for 15 min at 37°C. Then, cells were washed and further incubated for 72 h. ■, pierisin-1 only. ●, pierisin-1 plus membrane fraction from HeLa cells. □, pierisin-1 plus membrane fraction from MEB4 cells. Each experiment was carried out twice, and an average value of proliferation activity is given.

Discussion

In the present study, we investigated the incorporation and cytotoxicity of pierisin-1, a putative ADP-ribosylating toxin. Direct observation of HeLa cells incubated with Cy2 fluorescent dye-conjugated pierisin-1 clearly demonstrated that the protein bound to the cell surface and was then internalized into these cells. Moreover, 1–233 N-terminal polypeptide and 234–850 C-terminal polypeptide were not cytotoxic to HeLa cells when presented separately but exhibited cytotoxic effects when presented together in the culture medium. Such cytotoxic activity was increased about four times when the mixture of N- and C-terminal polypeptides were preincubated for 1 h at 25°C and added to the culture medium. Therefore, association of the N- and C-terminal fragments may produce a form accessible for binding to the receptor in the cell membrane. These observations suggest that the N- and C-terminal regions have distinct roles in the cytotoxic effects of pierisin-1, as in the case of other ADP-ribosylating toxins, including cholera and diphtheria toxins (14). No two-polypeptide combinations other than 1–233 plus 234–850 showed cytotoxic activity. Therefore, cleavage can occur only at the region around Arg-233–Ser-234 when cytotoxic activity is to be retained.

ADP-ribosyltransferase subunit or domain in the bacterial ADP-ribosylating toxins plays a key role in exhibiting their biological functions in target cells. The coding region of the A-subunit of diphtheria toxin is widely used as a negative selection marker for gene-targeting vector (15, 16). Introduction of Clostridium botulinum C3 exoenzyme into cells by microinjection and electroporation is also a common procedure for ADP-ribosylation of Rho protein in the cells (17–19). In addition, microinjected cholera toxin A-subunit has been reported to affect the signal transduction of the eggs at fertilization by ADP-ribosylating the GTP-binding proteins (20). In the case of pierisin-1, the sequence similarity of the N-terminal region to ADP-ribosyltransferases implies that this region has a direct role in the cytotoxic action of the toxin. As expected, the N-terminal polypeptide, but not the C-terminal or mutated N-terminal polypeptide, exhibited cytotoxicity and apoptosis-inducing activity in HeLa cells when it was introduced by electroporation and microinjection. These results indicate that, after incorporation into cells, the N-terminal region of pierisin-1 is responsible for cytotoxicity of the protein, probably mediated through its ADP-ribosylation activity.

The presence of the C-terminal polypeptide could be replaced by direct electroporation of the N-terminal polypeptide to achieve expression of cytotoxicity. This finding supports the notion that the C-terminal region of pierisin-1 plays a role in the incorporation of the N-terminal region. In fact, the N-terminal polypeptide could be incorporated only in the presence of the C-terminal polypeptide. Other bacterial toxins have also demonstrated that subunits or regions other than the ADP-ribosyltransferase domain play essential roles in the incorporation of the toxins into target cells by receptor binding (21–23). Therefore, C-terminal region of pierisin-1 might bind to a putative receptor for pierisin-1. Moreover, HA-33, being partially homologous to the C-terminal region of pierisin-1, is reported to bind to sialic acid or a galactose moiety (10, 11). These observations suggest that glycolipid or glycoprotein has a role by binding to pierisin-1. Consistent with this, the lipid fraction of the cell membrane in HeLa cells inhibited the cytotoxicity of pierisin-1. Thus, the glycolipid in the membrane might be a possible candidate for the receptor of pierisin-1. On the other hand, the electroporated N-terminal polypeptide of pierisin-1 was cytotoxic against MEB4 cells, which were resistant to pierisin-1 and hardly incorporated Cy2-pierisin-1. In addition, neither the membrane fraction nor total lipid fraction from MEB4 cells affected the cytotoxic activity of pierisin-1 against HeLa cells. These observations suggest that the resistance of MEB4 cells to pierisin-1 is probably because of the lack of a receptor for this toxin on the cell surface.

Proteolytic cleavage of the catalytic domains of diphtheria toxin and Pseudomonas exotoxin, both of which consist of a single polypeptide like pierisin-1, results in activation of these proteins (24, 25). Pierisin-1 is also likely to be cleaved by protease in mammalian cells, as its cytotoxic effects were potentiated by trypsin and generation of the N-terminal fragment from the intact protein in HeLa cells was observed. Furin is the cellular activating enzyme of diphtheria toxin and Pseudomonas exotoxin (24–26). This proteolytic enzyme belongs to the family of proprotein convertases (27, 28) and requires the sequence Arg-Xaa-Xaa-Arg for efficient cleavage of substrate (29–31). Interestingly, the amino acid sequence around the protease-sensitive region of pierisin-1 is Arg-227–Asp–Gln–Arg–Ser–Glu–Arg-233. Therefore, furin could be a possible enzyme to cleave pierisin-1 at the Arg-230–Ser-231 and Arg-233–Ser-234 sites.

On the basis of the above observations, a possible model for exertion of the cytotoxicity of pierisin-1 against mammalian cells is proposed as follows. Pierisin-1 binds to the receptor in the cell membrane by interaction of its C-terminal region. The entire protein is then internalized into the cell and cleaved into the N- and C-terminal fragments with intracellular protease. Finally, N-terminal fragment catalyzes the transfer of the ADP-ribose moiety of NAD to a target factor to induce cytotoxicity.

Our preliminary data of immunohistochemical analysis with anti-pierisin-1 antibody suggest that the protein is located in the fat body of the larvae of P. rapae. Because pierisin-1 may play roles in the pupation of P. rapae, it is possible that pierisin-1 selectively kills some types of larval cells, which may display receptors for the protein, whereas cells of the imaginal disk may not possess the receptor and thereby escape from the action of pierisin-1. On the other hand, pierisin-1 showing a strong cytotoxicity might contribute as a protective agent against invading organisms. Identification of the pierisin-1 receptor and the target molecule for ADP-ribosylation by pierisin-1 is very important for our understanding of the cytotoxic mechanism of the protein against mammalian cells and the actions of pierisin-1 in the insect.