Growth inhibition of mammalian cells by eosinophil cationic protein Takashi Maeda 1 , Midori Kitazoe 1 , Hiroko Tada 1 , Rafael de Llorens 2 , David S. Salomon 3 , Masakazu Ueda 4 , Hidenori Yamada 1 and Masaharu Seno 1 1 Department of Bioscience and Biotechnology, Faculty of Engineering, Graduate School of Natural Science and Technology, Okayama University, Japan; 2 Department of Biology, Faculty of Sciences, University of Girona, Spain; 3 Tumor Growth Factor Section, Basic Research Laboratory, National Cancer Institute, National Institutes of Health, Bethesda, MD, USA; 4 Department of Surgery, Keio University School of Medicine, Tokyo, Japan Eosinophil cationic protein (ECP), one of the major components of basic granules of eosinophils, is c ytotoxic to tracheal epithelium. However, the extent of this eect on other cell types has not been evaluated in vitro.Inthis study, we evaluated the eect of ECP on 13 mammalian cell lines. ECP inhibited the growth of several cell lines including those derived from carcinoma and leukemia in a dose-dependent manner. The IC 50 values on A431 cells, MDA-MB-453 cells, HL-60 cells and K562 cells were esti- matedtobe% 1±5 l M . ECP signi®cantly suppressed the size of colonies of A431 cells, and decreased K562 cells in G 1 /G 0 phase. However, there was little evidence that ECP killed cells in either cell line. These eects of ECP we re no t enhanced by extending its N-terminus. Rhodamine B isothiocyanate-labeled ECP started to bind to A431 cells after 0.5 h and accumulated for up to 24 h, indicating that speci®c anity for the cell surface may be important. The anity of ECP for heparin was assessed and found to be reduced when tryptophan residues, one of which is located at a position i n the catalytic subsite o f ribonuclease in ECP, were modi®ed. The g rowth-inhibitory eect was also attenuated by this modi®cation. These results suggest that growth inhibition by ECP is dependent on cell type and is cytostatic. Keywords: cell cycle; colony formation; cytostatic eect; eosinophil cationic protein (ECP); growth inhibition. Eosinophil cationic protein (ECP) is one of the major components of eosinophilic granules with a molecular mass ranging from 16 to 21.4 kDa. It exhibits various biological effects both in vitro and in vivo [1,2]. It is classi®ed as a member of the ribonuclease (RNase) A supergene family because of homology of both nucleotide and amino-acid sequences. The homology of amino-acid sequences between human ECP and human RNase 1 is % 30% [3,4]. On the other hand, ECP s hows signi®cant sequence homology (70%) with eosinophil-derived neurotoxin (EDN), which is another human RNase and a component of basic granules in eosinophils [5]. Recently t he 3D structure of E CP has been determined and con®rmed the similarity of its s tructure to other members of the pancreatic-type RNases [6,7]. Some substitutions of amino-acid residues in the catalytic subsites are consistent with the weak RNase activity of ECP. ECP is % 100±2000-fold less active than EDN depending on the type of substrate [8,9]. ECP is bactericidal [10], helminthotoxic [11±17], elicits the Gordon phenomenon when injected intrathecally into rabbits [18,19], and is cytotoxic to tracheal epithelium [20,21]. Although the mechanism o f its cytotoxicity is not completely understood, it is suggested to be due to the pore- forming activity of ECP, which destabilizes lipid membranes [22] a nd is unrelated to its RNase activity [14,23]. This is consistent with data showing that the cytotoxicity of ECP is greater than th at of E DN [13,19,24 ]. In this study, we have assessed the effect of ECP on the growth of 13 mammalian cell lines. The results s how that ECP is g rowth inhibitory depending on the cell t ype and is cytostatic but not cytotoxic. Fluorescent labeled ECP is shown to enter the c ell whereas RNase A does not. A speci®c af®nity for the cell surface may b e part of its cytostatic effect. This ability of ECP to bind to the cell surface is also shown t o depend on t ryptophan residues. MATERIALS AND METHODS Cell cultures Rat aortic smooth muscle A10 cells, human epidermoid carcinoma A431 cells, squamous carcinoma TE-8 cells derived from human esophageal cancer, HC-11 cells cloned from normal mouse mammary gland epithelia, and mouse metastatic melanoma-derived B16-BL6 cells were maintained as described previously [25±27]. Simian virus (SV)-40-transformed Balb/c 3T3 cell line SV-T2 [28], SV-40 transformed mouse Swiss/3T3 ®broblast cell line 3 T3-SV40 [29], mouse cell line LL/2 established as Lewis lung carcinoma [30], human colorectal adenocarcinoma cell line HT-29 [31], human chronic myelogenous leukemia cell line K562 [32], human acute promyelocytic leukemia cell line HL-60 [33], and human breast cancer cell line s MDA-MB-453 [34] and Correspondence to M. Seno, Department of Bioscience and Biotech- nology, Faculty of Engineering, Graduate school of Natural Science and Technology, Okayama University, 3.1.1 Tsushima-Naka, Okayama 700-8530, Japan. Fax/Tel.: + 81 86 251 8216, E-mail: senom@biotech.okayama-u.ac.jp Abbreviations: ECP, eosinophil cationic protein; EDN, eosinophil- derived neurotoxin; RNase, ribonuclease; SV, simian virus; MTT, 3-(4,5-dimethylthiazol-2-yl)-diphenyltetrazolium bromide; RITC, rhodamine B isothiocyanate; NBS, N-bromosuccinimide. (Received 31 August 2001, revised 26 October 2001, accepted 5 November 2001) Eur. J. Biochem. 269, 307±316 (2002) Ó FEBS 2002 T-47D [35] were obtained f rom American Type Culture Collection (USA) or Dainippon-Pharmaceutical Co. ( Japan) and maintained as directed. Preparation of recombinant human ECP HumanECPcDNAwasisolatedandexpressedinan Escherichia coli T7 expression system as described previ- ously [7]. Computer analyses fo r the sec retion signal [36,37] predicted the cleavage site of ECP to be between Gly23 and Ser24. On the other hand, ECP puri®ed fro m normal human eosinophils had Arg28 at its N-terminus [38]. Hence two different types of ECP were prepared. To distinguish between them, the one with an N-terminal extension from Ser24 was designated (À4) ECP. Puri®ed ECP and (À4) ECP w ere assessed for RNase activity on yeast RNA by the perchloric acid precipitation method as previously described [39] and for bactericidal activity against Staphy- lococcus aureus 209P FDA by counting the colonies on plates [23]. The N-terminal sequences, CD spectra, and apparent molecular masses in SDS/PAGE of both proteins were con®rmed to be as designed except for the ®rst methionine residue of (À4)ECP,whichwasprocessedoff. MTT assay for cell growth The effect of ECP on the growth of various cell lines was assessed by colorimetric assay using 3-(4,5-dimethylthiazol- 2-yl)-diphenyl-tetrazolium b romide (MTT) [25]. Cells were plated into 96-well plates (Nalge-Nunc, USA) in appropri- ate media containing 10% fetal bovine serum at 500 cells per well. After 24 h, each sample was added a t the indicated concentration (0±10 l M ). Four days after plating, the medium was replaced with fresh medium containing each sample at the same concentration. After a further 4 days of cultivation, MTT (5 mgÁmL À1 in NaCl/P i ) was added, and cell growth was monitored by measuring A 570 . Counting of viable cells The number of K562 cells under v arious conditions was counted. First, 25 000 cells were seeded into a 35-mm dish (Falcon), and appropriate concentrations of ECP (À4) or ECP and RNase A were simultaneously add ed. After 1±3 days of cultivation, viable cells unstained with Trypan blue were counted with a hemocytometer. Observation of cell morphology K562 and A 431 cells were seeded into 24-well plates or 35-mm dishes, and, after 24 h, ECP was added at a concen- tration of 10 l M . At appropriate times during the culture, the morphology of the cells was observed with a phase- contrast inverted microscope (CK -2; Olympus, Tokyo, Japan) equipped with a charge-coupled device video camera. Cell cycle analysis K562 cells were seeded at 500 000 cells per 6 0-mm dish in the growth medium. After 2 4 h, the medium was changed to fresh medium with 10 l M ECP or bovine RNase A (Sigma). After 3 days of treatment, cells were harvested, washed with NaCl/P i , treated with NaCl/P i containing 0.25% Triton X-100 and 0.15 mgÁmL À1 RNase A for 15 m in at room temperature. The cells were then ®xed in 70% ethanol overnight at 4 °C and this was followed by further treatment with RNase A (0.1 mgÁmL À1 )inNaCl/P i for 10 min at 37 °C. The DNA o f the ®xed cells was stained with propidium iodide (50 mgÁmL À1 ) for 30 min at room temperature, and the cells were analyzed by FACSCalibur (Becton Dickinson). Assay of colony formation A431 cells suspended in the medium were seeded into 35-mm dishes at 10 000 cells per dish. After 24 h, ECP was added at 10 l M to the medium and the cells were cultured for an additional 3 days. The medium containing ECP was then changed. Seven days afte r seeding, the cells were ®xed with 10% formaldehyde a nd stained with C rystal violet. The number of colonies was counted, a nd the area occup ied by the colonies was evaluated by image scanning assisted by a computer. Fluorescence microscopy ECP and bovine RNase A were labeled with rhodamine B isothiocyanate (RITC; Sigma) as previously described [40]. Cells were seeded into an eight-well Laboratory-Tek Chamber Slide (Nunc) at 20 000 cells per well and cultured. After 24 h, cells in each well were treated with RITC-labe led protein at a concentration of 1 l M for 0.5±24 h, then washed with NaCl/P i , and observed under a ¯uorescent microscope (BX40; Olympus). Hoechst 33342 dissolved in NaCl/P i to 2 l M (Molecular Probes) was used for nuclear staining. Oxidation of tryptophan residues Two tryptophan residues were modi®ed by oxidation with N-bromosuccinimide (NBS; Sigma) as p reviously described [41]. Brie¯y, 1.37 mg ÁmL À1 NBS dissolved in 50 m M sodium acetate, pH 4.5, was gradually added to a solution of ECP (1.6 mgÁmL À1 ) in the same buffer. Oxidation was monitored by measuring the decrease in A 280 during t he course of the reaction. After dialysis against Milli-Q water, the solution of modi®ed ECP was assessed for amino-acid composition and presence of tryptophan residues. Heparin-af®nity column chromatography ECP with or without modi®cation was applied to a heparin af®nity column (Heparin-Cellulo®ne; 4 ´ 150 mm; Chisso, Japan) equilibrated with 50 m M phosphate buffer, pH 7.0, and eluted with a linear gradient of NaCl (0.2±0.7 M per 60 min) at a ¯ow rate of 0.6 mLÁmin À1 .TheA 273 derived from tyrosine residues w as monitored and the af®nity of each protein for heparin was evaluated as the retention time of the peak t op of each pro®le. RESULTS ECPs with two different N-termini As there is a discrepancy in the N-terminus between the predicted and the puri®ed ECP protein, post-translational 308 T. Maeda et al.(Eur. J. Biochem. 269) Ó FEBS 2002 processing might be involved in the truncation of t he N-terminal sequence of ECP. We thought that it was important to assess the effect of this N-terminal extension in ECP because EDN has a similar four amino-acid extended form that conferred cytotoxic activity on KS Y-1 cells, which are neoplastic endothelial cells derived f rom Kaposi's sarcoma [42]. We therefore expressed two types of recom- binant human E CP using the T7 expression system (Fig. 1A). RNase activity of ECP and (À4) ECP against yeast RNA was 100 times lower than that of bovine RNase A (Fig. 1B), which is consistent with a previous report on the activity of ECP puri®ed from eosinophils [43]. ECP and (À4) ECP showed no difference in RNase activity (Fig. 1 B). Both forms of ECP exhibited bactericidal activity against S. aureus whereas RNase A did not show any activity (Fig. 1C). Therefore, both forms of recombinant ECPs were biologically active. Effect of ECPs on various cell lines The growth-modulatory e ffects of ECP and (À4) ECP were assessed on 1 3 cell lines derived from humans and rodents. The results are summarized in Table 1. ECP showed the strongest inhibition of growth in leukemia-derived cells K562 and HL-60 with an IC 50 (concentration that causes 50% inhibition) of 1.1 l M . A431, M DA-MB-453 and HC-11 cells were also sensitive to ECP with IC 50 values of % 4±6 l M .The(À4) ECP protein also inhibited the growth Fig. 1. N-Terminal sequences of ECP, its ribonucleolytic activity and bactericidal activity. (A) The signal cleavage site is predicted to lie between Gly23 and Ser24 whereas the N-terminus is reported to be Arg28 w hen puri®ed from eosinophils. R ecombinant E CP was prepared as the mature form starting from Arg28 as indicated by the arrow at the top, and (À4) ECP was prepared with an extension of four amino-acid residues (in bold letters) as also indicated by the arrow at the bottom. RNase activity against yeast RNA (B) and bactericidal activity against S. aureus (C) of each ECP (d)and(À4) ECP (j)wereeval- uated. Bovine RNase A ( s)isacontrol. Table 1. E ect of ECP and (À4) ECP on cell growth. All assays were carried out in quadruplicate in a 96-well plate and SD was calculated. SC, Percentage of cells that survived at 10 l M ECP or (À4) ECP; NT , not test ed; NA, no t applicable . Cell line Origin ECP (À4) ECP IC 50 (l M ) SC (%) IC 50 (l M ) SC (%) Human K562 Chronic myelogenous leukemia 1.1 1.7  0.3 2.0 3.0  1.3 HL-60 Acute promyelocytic leukemia 1.1 2.1  0.5 2.0 5.4  2.4 A431 Epidermoid carcinoma 4.0 38.7  7.6 6.0 49.3  9.7 MDA-MB-453 Breast carcinoma (mammary gland) 4.0 31.3  4.7 NT NT TE-8 Squamous carcinoma NA 68.5  2.8 NT NT T-47D Ductal carcinoma (mammary gland) NA 90.3  3.9 NT NT HT-29 Colon adenocarcinoma NA 75.3  4.7 NA 80.5  2.8 Mouse B16-BL6 Metastatic melanoma NA 73.5 + 9.0 NT NT LL/2 Lewis lung carcinoma NA 93.5  5.5 NA 100.6  4.5 3T3-SV40 SV-40-transformed Swiss 3T3 cells NA 57.6  9.8 NA 58.9  8.5 SV-T2 SV-40-transformed Balb/c 3T3 cells NA 90.3  8.7 NA 95.2  8.2 HC-11 Normal mammary gland epithelial cells 6.0 42.8  5.2 6.0 38.3 + 4.4 Rat A10 Normal aortic smooth muscle cells NA 84.5 + 2.3 NT NT Ó FEBS 2002 Cytostatic eect of ECP (Eur. J. Biochem. 269) 309 of these cell lines. However (À4) ECP was less active than ECP. T-47D, LL/2 and SV-T2 cells were resistant to ECP while TE-8, HT-29, B16-BL6, 3T3-SV40 and A10 cells were marginally sensitive such that t he IC 50 values could not be accurately calculated but were more than 10 l M . Effect of ECP on A431 and K562 cells At 5 l M ECP, the grow th of K 562 cells was completely suppressed whereas growth inhibition of A431 cells was % 50% (Fig. 2). As ECP is a member of the supergene family of pancreatic-type RNases and is unique in this family for its basic pI, its growth-inhibitory effect was compared w ith that of both bovine R Nase A and poly( L -lysine) (a verage molecular mass  2900). Neither RNase A nor poly ( L -lysine) had any effect on the growth of thes e cel ls (Fig. 2). After 7 days of incubation, ECP and (À4) ECP appeared to inhibit the aggregation of K562 cells and to keep them sparse (Fig 2C,D) in contrast with control cells or cells treated with RNase A, which grew as aggregates (Fig. 2 A,B). It is interesting to note that even (À4) ECP, which a llowed cell growth because of its weaker effect (Fig. 3A), appeared to suppress cell aggregation. A431 cells are epithelial-like cells and have a typical cobblestone appearance (Fig. 2E). They show this cobblestone pattern even when seeded at lower density, as shown in Fig. 4A for instance. Although this morphology was not affected by RNase A (Fig. 2F), the cells treated with ECP for 5 days were more stellate in appearance (Fig. 2G). These cells resumed growth when ECP w as removed from the culture medium (Fig. 2H). As the growth inhibition of K562 cells in Fig. 2 was shown after 7 days of treatment with ECP, which was Fig. 2. Suppression of cell growth in the presence of ECP. Left, the percentages of viable cells under various concentrations of RNase A (s), po ly ( L -lysine) (n), (À4) ECP (j)andECP(d) were plotted. Growth of K562 and A431 cells was monitored by MTT assay. Each assay was carried out in quadruplicate and standard deviation was calculated and depicted in each vertical line. Right, K562 cells seeded at 500 cells per 35-mm dish were cultured for 7 days in the regular medium (A) and in the presence of 10 l M each RNase A (B), ECP (C) and (À4) ECP (D). A431 cells seeded at 500 cells per 35-mm dish and cultured for 5 days in the regular medium (E) and in the presence of 10 l M each RNase A (F) and ECP (G). Four days after p lating, the medium was replaced with fresh medium containing each sample at the same concentration. ECP-treated A 431 cells (G) were further cultured f or 3 days in the regular medium without ECP (H). (G) and (H) sh ow the same ®eld of t he same dish . Original magni®cations of the plates are ´ 10. 310 T. Maeda et al.(Eur. J. Biochem. 269) Ó FEBS 2002 supplemented by the medium change, this effect was assessed in the ®rst 3 days of treatment by monito ring the cell number (Fig. 3A). (À4) ECP was less active as a growth inhibitor during this p eriod and the effect of ECP lasted almost throughout not allowing any appreciable increase in cell number. K562 cells treated with ECP were further analyzed by ¯ow cytometry; a signi®cant decrease was observed in the population of cells in the G 1 /G 0 phase of the cell cycle, and a small increase in the dying population when compared with K562 cells culture d in regular medium or cells treated with RNase A (Fig. 3B). However, the popu- lation in the G 2 /M phase of the cell cycle was not altered, and the total cell number was unaffected. A small number of K562 cells in the G 1 /G 0 phase of the cell c ycle that had been treated with E CP or (À4)ECPappearedtobedead. The effect of ECP in the early period without medium change was monitored on A431 cells (Fig. 4). Up to 4 days after the addition of ECP, A431 cells were more sparse and were more ®broblastic in appearance (Figs 4D,E) in con- trast with control A431 cells (Fig. 4A,B) or A431 cells treated with RNase A (Fig 4G,H). In the presence of ECP, the cells were more ¯at and spread out after 6 d ays of treatment (Fig. 4F). Nuclei were more pronounced because of the low density and ¯attened shape of the cells (Figs 4A± C). Very recently, we found that ECP-treated Balb/c 3T3 cells exhibited a similar change in morphology with enhanced expression of vinculin (M. Kitazoe, T. Maeda, H. Tad a, R. de Llorens, D . S . S alomon, M. Ueda, H. Yamada & M. Seno, unpublished results). The e ffect of ECP o n the cell shape might be due to the regulation of vinculin gene exp ression as previously described [44,45]. The numbers of A431 colonies in the dishes (2500  110 colonies per 35-mm dish) that received ECP or RNase A were almost equivalent to the number of colonies in control A431 cells (Fig. 5A). However, ECP produced a signi®cant decrease in the size of the colonies of 60% compared with control cells (Fig. 5B). These results d emonstrate t hat E CP impairs the growth of cells. Cellular localization of ECP To assess whether E CP could b e localized, A431 cells were incubated with RITC-labeled ECP at 37 °C for various times (Fig. 6). From 0.5 to 3 h, A431 cells exhibited increased levels of ¯uorescence labeling in the cytoplasm rather than the nuclei. After 24 h, the ¯uorescence increased in the cells. There was no uptake of RITC-labeled RNase A into the cells, indicating that ECP may interact prefe ren- tially with a receptor or binding protein on t he cell surface. We attempted to observe the speci®c binding using ECP and RITC-labeled ECP by monitoring the level of ¯uo rescence, but the change in ¯uorescence level caused by the compe- tition of ECP and RITC-labeled ECP could not be detected. This is probably because at least 1 l M RITC-labeled ECP was r equired t o detect the ¯uorescence, and this concentra- tion may be too high to compete with the ECP, the practical maximum c oncentration of w hich is 10 l M . Although we could not show speci®c binding of ECP to c ells using a competition a ssay, we could assess the af®nity of ECP for heparin using heparin af®nity co lumn chromatography (Fig. 7 A). ECP was eluted at about 0.64 M NaCl, and (À4) ECP a nd NBS-modi®ed ECP were eluted at 0.60 and 0.56 M NaCl, respectively. The amino-acid composition of NBS-modi®ed ECP con®rmed that only tryptophan resi- dues had been modi®ed. As shown in F ig. 7B, these tryptophan residues a re located in the RNA c atalytic site of ECP and may contribute to the binding to heparin. The cleft of the catalytic site possibly functions as the site of attachment to proteoglycans on the cell s urface. As NBS- modi®ed ECP inhibited the growth of A431 cells less than (À4) ECP and the IC 50 could not be determined (T. Maeda, D. L. Newton, S. M. Rybak, unpublished results), the af®nity for heparin must also be responsible for the growth- inhibitory effect. DISCUSSION This is the ®rst report to demonstrate that ECP has a growth-inhibitory effect that is cytostatic and dependent on cell type. The growth of four of the seven human cell lines Fig. 3. Growth-inhibitory eect of ECP on K562 cells. (A) Cells were seeded at 25 000 cells per 35-mm dish. Simultaneously, 10 l M each RNase A, ECP and (À4) ECP were added t o the medium. After the indicated number of days of culture, the viable cells in the dishes were counted. The cell numbers are the average from three ind ependent experiments and standard deviations are depic ted by vertical lines on the top of each bar. The ho rizontal gray line shows the c ell n umber seeded at the beginning of the experiments. (B) Cells were seeded at 500 000 cells per 60-mm dish, cultured for 3 days in the presence or absence of 10 l M ECP or RNase A and analyzed by a ¯ow cytometer. The area of dead cells is shaded. Peaks I and II show the population of cells in G 1 /G 0 and G 2 /M phase, respectively. Ó FEBS 2002 Cytostatic eect of ECP (Eur. J. Biochem. 269) 311 Fig. 4. Time course change in the morphology of A431 cells treated with ECP. A431 cells were seeded into a 24-well plate at 5000 cells per well. After 24 h, medium was changed to a fresh one (A, B, C) or one containing 10 l M ECP (D, E, F) or RNase A. (G, H, I). Cells were photographed 1 day (A, D, G), 4 days (B, E, H) and 6 days (C, F, I) after the change of medium. Original magni®cations are ´ 10. Fig. 5. Growth-inhibitory eect of ECP on A431 cells. A431 cells were seeded at 5000 cells per 35-mm dish, and 2500  110 colonie s were formed in the presence or absence of 10 l M RNase A or ECP (A). The percentage of areas of colonies treated with RNase A and ECP were calculated taking colonies cultured in the grow th m edi um as 1 00% (B). This experiment was rep eated t hree times. Ph oto- graphs are the typical pattern of the colonies, and the percen tages are the mea ns of each result with the standard deviations within 10%. 312 T. Maeda et al.(Eur. J. Biochem. 269) Ó FEBS 2002 was inhibited, whereas the rodent cell lines were relatively resistant. The resistance of the rodent cell lines to the growth-inhibitory effects of human ECP may be due to the evolutionary divergence of ECP, which resulted in signi®- cantly low homology of e osinophil-derived RNase between species [46,47]. In this study, ECP suppressed the growth of K562, HL-60, A431 and MDA-MB-453 cell lines in an IC 50 range o f 1±4 l M . A lthough the primary structure of human ECP shows the closest identity (67%) with human EDN, the N-terminal extension of which confers cytotox- icity against KS Y-1 cells [42], the N-terminal extension of ECP did not produce any enhancement of the growth- inhibitory effects on this cell line (personal communi- cation, D. L. Newton and S. M. Rybak, National Cancer Institute, National Institutes of Health, Frederick, MD, USA). On the contrary, the N-terminal extension of E CP appears to impair the inhibito ry effect of ECP on some cell lines. Fig. 6. A431 cells treated with RITC-labeled ECP. A431 cells were seeded into the eight-well Laboratory-Tek chamber slide at 20 000 cells per well. After 24 h, 1 l M each RITC-labeled ECP and RNase A was added to the c ulture medium , and the cells we re ®xed and detected by ¯uorescent microscopy at the time indicated. RITC-labeled ECP or RNase A was visualized (top) and nuclei of the cells stained with Hoechst 33342 (bottom). The same ®eld was assigned in the same column at each time. The scale bar i s equivalent to 50 lm. Fig. 7. Heparin anity column chromatography of ECP (A) and schematic diagram of ECP depicted with RasMol v2.6 according to the PDB entry 1DVT (B). (A)ECPmodi®edwithNBS(a)(À4) ECP (b) or ECP (c) was applied to a heparin column and eluted with a linear gradient of NaCl. The retentiontimeofeachpeakofpro®lesis43min(0.64 M NaCl) for NBS-modi®ed ECP, 48 min (0.60 M NaCl) for (À4) ECP and 53 min (0.56 M NaCl) for ECP. RNase A passed through this anity column under the same conditions. (B) The backbone of peptide bonds is drawn in gray. The secondary-structure elements are helices and arrows for a he lic es and b strands, respectively. Two tryptophan residues, W10 and W35, and the other amino-acid residues in the catalytic subsites are in black with their side chains. H64 is located in the P0 subsite, H15, K38 and H128 in the P1 subsite, and W10 in the P2 subsite. Ó FEBS 2002 Cytostatic eect of ECP (Eur. J. Biochem. 269) 313 ECP is unique among RNase A enzymes because of its high arginine content. Therefore, the nonspeci®c effect of growth inhibition may be replicated by other polycations. To investigate this, we checked the effect of poly( L -lysine), which has a cationic charge that is almost equivalent to that of ECP. It did not have any growth-inhibitory effect on K562 and A431 cells (Fig. 2). Although we could not obtain polyarginine, we assessed the effect of the third helix of the Drosophila antennapedia homeodomain [48] and the basic region of HIV-Tat p rotein [49], both of w hich are rich in basic amino acids and known to enhance cellular uptake. Neither of these peptides showed any growth-inhibitory effect similar to ECP (data not shown). Therefore, the cationic charge could not replicate the effect of ECP. Recently, it has been shown that the amphipathic helix structure o f the basic region of Tat protein may be important for the uptake of the protein into cells [50]. This mechanism may not apply to ECP as ECP is an arginine- rich protein but does not have the cluster of arginine residues like Tat protein. On the other hand, highly cationized RNase molecules are cytotoxic but are not cytostatic like ECP [51]. The cytostatic growth-inhibitory effect of ECP was not suf®cient to induce cell death, as the initial number o f viable K562 cells did not signi®cantly decrease in the presence of ECP. Cell cycle a nalysis showed that ECP decreased K562 cells in the G 1 /G 0 phase without having an effect on the G 2 / M phase. The number of dying cells increased slightly but this was not signi®cant. The growth-in hibitory effect of ECP was reversible as A431 cells treated with ECP resumed normal g rowth w hen ECP was removed from the medium. In addition, ECP affected the size of colonies of A431 cells but not colony number. Furthermore, we found no effect of ECP on the frequency of apoptosis as assessed by DNA ladder formation and phosphatidylserine externalization (data not shown). RITC-labeled ECP appeared to accumulate in the cyto- plasm after 24 h. However, whether endocytosis of ECP ocurred is still unclear. It may be in vesicles or endosomes. As the cellular outlines were not enhanced during the course of the incubation, we conclud ed that speci®c accumulation of ECP o n t he cell surface was not occurring. The most probable explanation for this observation is the presence o f cell s ur- face receptors for ECP. We could no t show the presence of a speci®c high-af®nity binding site for ECP on cells by com- petition assay. This is probably because the af®nity of ECP for binding to extracellular matrix proteins such as heparan sulfate is extremely low and rapidly dissociates. However, it is very interesting to note that ECP can bind to heparin and that tryptophan residues contribute to this af®nity. As tryp- tophan residues are also responsible for the af®nity for galactose and lactose residues of some lectins [52], it i s feasible that this is also the case for ECP. The trypto phan residues are Trp10 and/or Trp35 in ECP. Interestingly, Trp10 is located at the P2 subsite of the catalytic domain of RNase and controls the weak RNase activity of the p rotein [6]. Trp35 is a unique amino-acid residue at a position in Loop-3 of the RNases. As both tryptophan residues are located on t he same side of the molecule, it is dif®cult to ascertain which residue is more critical for the the binding of ECP to heparin or other carbohydrates. In conclusion, ECP has a cell-type-speci®c cytostatic growth-inhibitory effect. It is possible that this activity is due to the presence of the tryptophan residues. We are now producing mutant ECP proteins to assess this aspect. Further evaluation of the ECP molecule including the generation of a mutant protein will enable us to test the effects of ECP on in¯ammatory diseases, in w hich uncon- trolled cell growth c ould contribute t o a delay in wound healing. In addition, strong local in¯ammatory responses that are speci®c to cytokines such as interleukin-4, are capable in some cases of mediating regression of tumors [53,54]. In¯ammatory in®ltrates comprised of eosinophils may also play an important role at the primary site of tumor regression by releasing ECP as part of the cascade of inducing a tumor-speci®c T-cell response. Optimization of the biological activity of the ECP p rotein to target the cell surface could bring IC 50 values down t o practical levels so that it might be used as an anti-cancer reagent in the same way as others [27,55±57]. ACKNOWLEDGEMENTS We thank Drs S. Rybak and D. Newton for assaying the eects of ECP and (À4) ECP on KS Y-1 cells, Professor K. O guma for providing S. aureus cells, Professor M. Hikida for help with ¯ow-cytometric analysis, and Drs R. Sasada, S. Ishikami and J. Futami for helpful discussions and suggestio ns throughout this work. T his work was partly supported b y a grant-in-aid f rom the Ministry of Education, Science , Culture and Sports of Japan. R. de Ll. was supported by the Spanish Ministerio de Educacio  n y Cultura (grant SAF 9 8-0086 and 2FD 97- 0872) and by Generalitat de Catalunya (grant SGR97-240). REFERENCES 1. Rosenberg, H.F. (1998) The eosinophil ribonucleases. Cell. Mol. Life Sci. 54, 795±803. 2. Giembycz, M.A. & Lindsay, M.A. (1999) Pharmacology of the eosinophil. Pharmacol. Rev. 51, 213±340. 3. Seno, M., Futami, J., Kosaka, M., Seno, S. & Yamada, H. (1994) Nucleotide sequence encoding human pancreatic ribonucle ase. Biochim. Biophys. Acta 1218, 4 66±468. 4.Futami,J.,Tsushima,Y.,Murato,Y.,Tada,H.,Sasaki,J., Seno, M. & Yamada, H. (1997) Tissue-speci®c expression of pancreatic-type RNases and RNase inhibitor in humans. DNA Cell Biol. 16, 413±419. 5. Gleich, G.J., Loegering, D.A., Bell, M.P., Checkel, J.L., Acker- man, S.J. & McKean, D.J. (1 986 ) Biochemical and function al similarities between human eosinophil-derived neurotoxin and eosinophil cationic protein: homology with ribonuclease. Proc. Natl Acad. Sci. USA 83, 3146±3150. 6. Boix, E., Leonidas, D.D., Nikolovski, Z., Nogues, M.V., Cuchillo, C.M. & Acharya, K .R. ( 1999) Crystal structure of eosinophil cationic protein at 2.4 A Ê resolut io n. Biochemistry 38, 16794±16801. 7. Mallorqui-Fernandez, G., Pous, J., Peracaula, R., Aymami, J., Maeda,T.,Tada,H.,Yamada,H.,Seno,M.,deLlorens,R., Gomis-Ruth, F.X. & Coll, M. (2000) Three-dimensional crystal structure of human eosinophil cationic protein (RNase 3) at 1.75 A Ê resolution. J. Mol. Biol. 300, 1297±1307. 8. Slifman, N.R., Loegering, D.A., McKean, D.J. & Gleich, G.J. (1986) Ribonuclease activity associated with human eosinophil- derived neurotoxin and eosinophil cationic protein. J. Immunol. 137, 2913±2917. 9. Rosenberg, H.F. & Dyer, K.D. (1997) Diversity among the pri- mate eosinoph il-derived neurotoxin genes: a speci®c C-terminal sequence is necessary for enhanced ribonuclease activity. Nucleic Acids Res. 25, 3532±3536. 10. Lehrer, R.I., Szklarek, D., Barton, A., Ganz, T., Hamann, K.J. & Gleich, G.J. (1989) Antibacterial properties of eosinophil major 314 T. Maeda et al.(Eur. J. Biochem. 269) Ó FEBS 2002 basic protein and eosinophil cationic p rotein. J. Immunol. 142, 4428±4434. 11. Ackerman, S.J., Gleich, G.J., Loegering, D.A., Richardson, B.A. & Butterworth, A.E. (1985) Comparative toxicity of puri®ed human eosinophil granule cationic proteins for sc histosomula of Schistosoma mansoni. Am. J. Trop. Med. Hyg. 34, 735±745. 12. Hamann, K.J., Gleich, G.J., Checkel, J.L., Loegering, D.A., McCall, J.W. & Barker, R.L. (1990) In vitro killing of micro®lariae of Brugia pahangi and Brugia malayi by e osinophil g ranule pro - teins. J. Immunol. 144, 3166±3173. 13. Hamann, K.J., Barker, R .L., Loegering, D.A. & Gleich, G.J. (1987) Comparative toxicity of puri®ed human eosinophil granule proteins for ne wborn larvae of Trichinella spiralis. J. Parasitol. 73, 523±529. 14. Molina, H.A., Kierszenbaum, F., Hamann, K.J. & Gleich, G.J. (1988) Toxic eects produced or m ediated by human eosinophil granule com ponents on Trypanosoma cruzi. Am. J. Trop. Med. Hyg. 38, 327±334. 15. McLaren, D.J., McKean, J.R., Olsson, I., Venges, P. & Kay, A.B. (1981) Morphological studies on the killing of schistosomula of Schistosoma mansoni by human eosinophil and neutrophil cationic proteins in vitr o. Parasite Immunol. 3, 359±373. 16. Yazdanbakhsh, M., Tai, P.C., Spry, C.J., Gleich, G.J. & Roos, D. (1987) Synergism between eosinophil cationic protein and oxygen metabolites in killing of schistosomula of Schistosoma mansoni. J. Immunol. 138, 3443±3447. 17. Waters, L.S., Taverne, J., Tai, P.C., Spry, C.J., Targett, G.A. & Playfair, J.H. (1987) Killing of Plasmodium falciparum by eosin - ophil secretory products. Infect. Immun. 55, 877±881. 18. Durack,D.T.,Sumi,S.M.&Klebano,S.J.(1979)Neurotoxicity of human eosinophils. Proc. Natl Acad. Sci. USA 76, 1443±1447. 19. Fredens, K., Dahl, R. & Venge, P. (1982) The Gordon phenom- enon induced by the eosinophil cationic protein and eosinophil protein X. J. Allergy Clin. Immunol. 70, 361±366. 20. Motojima, S., Frigas, E., Loegering, D.A. & Gleich, G.J. (1989) Toxicity of eosinophil cationic pro teins for guinea p ig tracheal epithelium in vitro. Am. Rev. Respir. Dis. 139, 801±805. 21. Fredens, K., Dybdahl, H., Dahl, R. & Baandrup, U. (1988) Extracellular deposit of the c ationic proteins ECP and EPX in tissue in®ltrations of eosinophils related to tissue damage. APMIS 96, 711±719. 22. Young, J.D., Peterson, C.G., Venge, P. & Cohn, Z.A. (1986) Mechanism of membrane damage mediated by human eosinophil cationic protein. Nature (London) 321, 613±616. 23. Rosenberg, H.F. (1995) Recombinant human eosinophil cationic protein. Ribonuclease activity is n ot essential fo r cytotoxicity. J. Biol. Chem. 270, 7876±7881. 24. Barker, R.L., Ten Loegering, D.A.R.M., Hamann, K.J., Pease, L.R. & Gleich, G.J. (1989) Eosinophil cationic protein cDNA. Comparison with other toxic cationic proteins and ribonucleases. J. Immunol. 143, 952±955. 25. Seno, M., Tada, H., Kosaka, M., Sasada, R., Igarashi, K., Shing, Y.,Folkman,J.,Ueda,M.&Yamada,H.(1996)Humanbeta- cellulin, a member o f the EGF fa mily dominantly expressed in pancreas and small intestine, is fully active in a monomeric form. Growth Factors 13, 181±191. 26. Seno,M.,DeSantis,M.,Kannan,S.,Bianco,C.,Tada,H.,Kim, N., Kosaka, M., Gullick, W.J., Yamada, H. & Salomon, D.S. (1998) Puri®cation and characterization of a recombinant human cripto-1 protein. Growth Factors 15, 215±229. 27. Futami,J.,Seno,M.,Ueda,M.,Tada,H.&Yamada,H.(1999) Inhibition of cell growth by a fused protein of human ribonuclease 1 and human b asic ®broblast growth factor. Protein E ng. 12, 1013±1019. 28. Aaronson, S.A. & Todaro, G.J. ( 1968) Development of 3 T3-like lines from Balb-c mouse em bryo cultures: transformation sus- ceptibility to SV40. J. Cell. Physiol. 72, 141±148. 29. O'Neill, F.J. (1975) Control of nuclear division in sv40 and ade- novirus type 12 transformed mouse 3t3 cells. Int. J. Cancer 15, 715±723. 30. Bertram, J.S. & Janik, P. (1980) Establishment of a cloned line of Lewis Lung Carc inoma cells adapted to cell cultu re. Cancer Lett. 11, 63±73. 31. von Kleist, S., Chany, E., Burtin, P., King, M. & Fogh, J. (1975) Immunohistology of the antigenic pattern of a continuous cell line from a human colon tumor. J. Natl Cancer Inst. 55, 555±560. 32. Andersson, L.C., Nilsson, K. & Gahmberg, C.G. (1979) K562: a human erythroleukemic cell line. Int. J. Cancer 23, 143±147. 33. Collins, S.J., Gallo, R.C. & Gallagher, R.E. (1977) Continuous growth and dierentiation of human myeloid leukaemic cells in suspension culture. Nature (London) 270, 347±349. 34. Cailleau, R., Olive, M. & Cruciger, Q.V. (1978) Long-term human breast carcinoma cell lines of metastatic origin: p reliminary char- acterization. In Vitro 14, 911±915. 35. Prager, A., Ben-Hur, E., Ch aitcik, S., Brenn er, H.J. & Riklis, E. (1981) Characterization of the response of a human breast carci- noma cell line (T-47D) to radiation and chemotherapeutic agents. Isr. J. Med. Sci. 17, 976±979. 36. Nakai, K. & Kanehisa, M. (1992) A k nowledge base for predict- ing protein localization sites in eukaryotic cells. Genomics 14, 897±911. 37. Nielsen, H., Engelbrecht, J., Brunak, S. & von Heijne, G. (1997) Identi®cation of prokaryotic and e ukaryotic signal peptide s and prediction of their cleavage sites. Protein Eng. 10, 1±6. 38. Peterson, C.G., Jornvall, H. & Venge, P. (1988) Puri®cation and characterization of eosinophil cationic protein from normal human eosinophils. Eur J. Haematol. 40, 415±423. 39. Futami, J., Seno, M., Kosaka, M., Tada, H., Se no, S. & Yamada, H. (1995) Recombinant human pancreatic ribonuclease produced in E. coli: importance of the amino-terminal sequence. Biochem. Biophys. Res. Commun. 216, 406±413. 40. Hiratsuka, T. (1987) Selective ¯uorescent labeling of the 50-, 26-, and 20-kilodalton heavy chain segments of myosin ATPase. J. Biochem. (Tokyo) 101, 1457±1462. 41. Spande, T.F., Green, N.M. & Witkop, B. (1966) The reactivity toward N-bromosuccinimide of tryptophan in enzymes, zymogens, and inhibited enzymes. Biochemistry 5, 1926±1933. 42. Newton, D.L. & Rybak, S.M. (1998) Unique recombinant human ribonuclease and inhibition of Kaposi's sarcoma cell growt h. J. Natl Cancer Inst. 90, 1787±1791. 43. Sorrentino, S. & Libonati, M. (1997) Structure±function rela- tionships in human ribonucleases: main distinctive features of the major RNase types. FEB S Lett. 404, 1±5. 44. Rodriguez F ernandez, J.L., Geiger, B., Salomon, D., Sabanay, I., Zoller, M. & Ben-Ze'ev, A. (1992) Suppression of tumorigenicity in transformed cells after transfection with vinculin cDNA. J. Cell Biol. 119, 427±438. 45. Rodriguez Fernandez, J.L., Geiger, B., Salomon, D. & Ben-Ze'ev, A. (1993) Suppression of vinculin expression by antisense trans- fection confers changes in cell m orphology, motility, and ancho- rage-dependent growth of 3T3 cells. J. Cell Biol. 122, 1285±1294. 46. Larson, K .A., Olson, E.V., Madden, B.J., Gleich, G.J., Lee, N.A. & L ee, J.J. (1996) Two highly homologous ribonuclease ge nes expressed i n mouse eosinophils identify a larger sub group of the mammalian ribonuclease superfamily. Proc. Natl Acad. Sci. USA 93, 12370±12375. 47. Singhania, N.A., Dye r, K.D., Zhang, J., Deming, M.S., Bonville, C.A., Domachowske, J.B. & Rosenberg, H.F. (1999) Rapid e vo- lution of the ribonuclease A supe rfamily: adaptive exp ansion of independent gene clusters in rats an d mice. J. Mol. Evol. 49,721± 728. 48. Derossi, D., Calvet, S., Trembleau, A., Brunissen, A., Chassaing, G. & Prochiantz, A. (1996) Cell internalization of the third helix of Ó FEBS 2002 Cytostatic eect of ECP (Eur. J. Biochem. 269) 315 the Ante nnapedia homeodomain is receptor-independent. J. Biol. Chem. 271, 18188±18193. 49. Schwarze, S.R., Ho, A., Vocero-Akbani, A. & Dowdy, S.F. (1999) In vivo protein transduction: delivery of a biologically active protein into the mouse. Science 285, 1569±1572. 50.Ho,A.,Schwarze,S.R.,Mermelstein,S.J.,Waksman,G.& Dowdy, S.F. (2001) Synthetic protein transduction domains: enhanced transduction potential in vitro andinvivo.Cancer Res. 61, 474±477. 51. Futami, J., M aeda, T., Kitazoe, M., Nukui, E., T ada, H., Seno,M.,Kosaka,M.&Yamada,H.(2001)Preparationof potent cytotoxic ribonucleases by cationization: enhanced cellular uptake and decreased interaction with ribonuclease inhibitor by ch emical modi®cation of carboxyl groups. Bio chemistry 40, 7518±7524. 52. Rini, J.M. (1995) Lectin structure. Annu. Rev. Biophys. Biomol. Struct. 24, 551±577. 53. Tepper, R.I., Coman, R.L. & Leder, P. (1992) An eosinophil- dependent mechanism for the antitumor eect of interleukin-4. Science 257, 548±551. 54. Pardoll, D.M. (1995) Paracrine cytokine adjuvants in cancer immunotherapy. Annu. Rev. Immunol. 13, 399±415. 55. Psarras, K., Ueda, M., Yamamura, T., Ozawa, S., Kitajima, M., Aiso, S., Komatsu, S. & Seno, M. (1998) Human pancreatic RNase1-human epidermal growth factor fusion: an entirely human Ôimmunotoxin analogÕ with cytotoxic properties against squamous cell carcinomas. Protein Eng. 11, 1285±1292. 56. Psarras, K., Ueda, M., Tanabe, M., Kitajima, M., Aiso, S., Komatsu, S. & Seno, M. (2 000) Ta rgeting activ ated ly mphocy tes with an entirely human immunotoxin analogue: human pancreatic RNase1-human IL-2 fusion. Cytokine 12 , 786±790. 57. Reiter, Y. & Pastan, I. (1998) Recombinant Fv immunotoxins and Fvfragmentsasnovelagentsforcancertherapyanddiagnosis. Trends Biotechnol. 16 , 513±520. 316 T. Maeda et al.(Eur. J. Biochem. 269) Ó FEBS 2002 . formation; cytostatic eect; eosinophil cationic protein (ECP); growth inhibition. Eosinophil cationic protein (ECP) is one of the major components of eosinophilic. also attenuated by this modi®cation. These results suggest that growth inhibition by ECP is dependent on cell type and is cytostatic. Keywords: cell cycle; colony formation;