Pore-forming peptide of pathogenic Entamoeba histolytica (amoebiasis/amoebapore/membranolytic peptides/peptide-peptide interaction)

Transcription

1 Pro. Nati. Aad. Si. USA Vol. 88, pp , September 1991 Medial Sienes Pore-forming peptide of pathogeni Entamoeba histolytia (amoebiasis/amoebapore/membranolyti peptides/peptide-peptide interation) MATTHIAS LEIPPE, SEBASTIAN EBEL, OEYVIND L. SCHOENBERGER, ROLF D. HORSTMANN, AND HANS J. MULLER-EBERHARD Department of Moleular Biology, Bernhard Noht Institute for Tropial Mediine, 2000 Hamburg 36, Federal Republi of Germany Contributed by Hans J. Muller-Eberhard, June 24, 1991 ABSTRACT A polypeptide that auses pore formation in target-ell membranes is impliated in the potent ytolyti ativity of pathogeni Entamoeba histolytia. Pore-forming material was purified to apparent homogeneity by a multistep proedure, and its analysis by NaDodSO4/PAGE revealed one peptide of 4-5 kda under nonreduing or under reduing onditions. Pore-forming ativity was measured by depolarization of liposome membrane potential and was found to be optimally expressed at low ph. Ative material preferentially inserted into negatively harged lipid vesiles. Treatment of purified amoeba peptide in solution or bound to liposomes with glutaraldehyde revealed oligomers upon NaDodSO4/PAGE, suggesting funtionally relevant peptide-peptide interations. The NH2-terminal amino aid sequene of the amoeba peptide was determined by protein sequening and revealed a strutural similarity to melittin, the membranolyti peptide of bee venom. Entamoeba histolytia, the protozoan parasite ausing amoebiasis, is apable of destroying the tissue of the infeted host and is ytolyti to a variety of ell types in vitro (1). Beause the ytolyti reation is ontat-dependent and takes plae within minutes of onjugation with target ells, it resembles that of ytotoxi lymphoytes (2). Lymphoytes lyse ells by sereting a protein that forms stable pores in the membrane of target ells (3). Suh pore-forming proteins have been found as mediators of membrane damage in a variety of organisms from prokaryotes to humans and inluding protozoan parasites (2-5). It has been reported (6, 7) that E. histolytia also produes a protein apable offorming pores in artifiial lipid bilayers and target-ell membranes and that this protein has a moleular mass of kda in its native state and of kda under denaturing and reduing onditions. The protein, named amoebapore, was partially purified reently from the 150,000 x g supernatant of amoeba lysates (8). Studies of the effet of amoeba lysates or partially purified material on planar lipid bilayers suggested that oligomerization of ative protomers ourred during formation of membrane pores (7-9). The present paper desribes the isolation of a polypeptide from pathogeni E. histolytia with pore-forming ativity, its NH2-terminal amino aid sequene, and its strutural similarity to melittin, the membranolyti polypeptide of bee venom. MATERIALS AND METHODS The following materials were purhased from Sigma: aprotinin (Mr 6500); rude phosphatidlyholine type 1I-S of soybeans; phosphatidylholine, dioleoyl; phosphatidylethanolamine type III of egg yolk; phosphatidylglyerol, dioleoyl; phosphatidylinositol of bovine liver; phosphatidylserine of The publiation osts of this artile were defrayed in part by page harge payment. This artile must therefore be hereby marked "advertisement" in aordane with 18 U.S.C solely to indiate this fat. bovine brain; sphingomyelin of egg yolk; trans-epoxysuinyl-l-leuylamido-(4-guanidino)butane (E-64); and valinomyin. 3,3'-Diethylthiodiarboyanine iodide was purhased from Kodak, fast protein liquid hromatographi olumns Mono Q HR 10/10, Superdex 75 pg 26/60, Superdex 200 pg 26/60, Superose 12 HR 10/30, and gel filtration alibration kits were from Pharmaia LKB, a fast protein liquid hromatographi hydroxylapatite olumn HPHT was from Bio- Rad, and protein moleular weight standards for NaDodSO4/ PAGE-rainbow markers-were from Amersham. E. histolytia Isolates. The pathogeni strain HM-1:IMSS of E. histolytia was ultured in axeni medium TYI-S-33 in plasti tissue ulture flasks (10). Trophozoites from ultures in late-logarithmi phase were harvested after being hilled on ie for 10 min. The amoebae were pelleted at 430 x g for 10 min at 4 C, washed twie in phosphate-buffered saline (PBS), and frozen at -70 C. Trophozoite Extrat. Frozen trophozoites (4.5 x 109 ells) were thawed in the presene of the proteinase inhibitors E-64 (10,M), benzamidine (6.4 mm), and phenylmethylsulfonyl fluoride (0.5 mm). The ells were frozen and thawed in two yles and entrifuged at 430 x g for 10 min at 4 C to remove nulei and debris. The supernatant was arefully removed and entrifuged at 150,000 x g for 40 min at 4 C. The 150,000 x g supernatant (60 ml) was stored at -70 C. Assay for Pore-Forming Ativity. Pore-forming ativity was determined essentially aording to Loew et al. (11). Liposomes were prepared as desribed by Pik (12) from rude soybean phosphatidylholine at a onentration of 40 mg-ml-1 in 50 mm K2SO4/0.5 mm EDTA/50 mm Tris maleate, ph 5.2. For the assay liposomes were diluted 1:4000 in the buffer used for the liposome preparation with K+ replaed by Na+ (uvette buffer). Addition of valinomyin to a final onentration of 1 nm resulted in a potassium diffusion potential that was monitored by the fluoresene quenhing of 3,3'-diethylthiodiarboyanine iodide dye (1,uM). Fluoresene was measured by a fluoresene spetrophotometer (model MPF-2A; Perkin-Elmer) using exitation and emission wavelengths of 620 nm and 670 nm, respetively. Poreforming ativity was measured as the initial hange in fluoresene intensity over time after adding the sample. One unit of ativity was defined as a fluoresene inrease to 5% of the prevalinomyin intensity in 1 min at 25 C. Absorption of Pore-Forming Ativity. Aliquots of trophozoite extrat (5,ul) were diluted with phosphate-buffered saline (15,l) and inubated for 3 hr at 4 C with phospholipid vesiles at various onentrations in uvette buffer (380,l). Vesiles were prepared by hydration of dried films of different phospholipids in onial-bottomed flasks with uvette buffer and subsequent sonifiation. After the inubation period 2 Al of the extrat/phospholipid mixture was assayed for ativity. Addition of vesiles alone did not hange fluoresene level of the assay. Purified amoeba peptide was adsorbed to phosphatidylinositol vesiles and subjeted to NaDodSO4/PAGE. 7659

2 7660 Medial Sienes: Leippe et al. Pro. Natl. Aad. Si. USA 88 (1991) Purifiation of Pore-Forming Peptide. All purifiation proedures were done at 40C, and samples were kept on ie. Buffers for the first two hromatographies ontained the proteinase inhibitors E-64 (10,uM) and benzamidine (6.4 mm). Plastiware with minimal adsorption properties (Mirosorp immuno tubes, Nun) was used. Dialysis and onentration proedures were kept to a minimum. Pore-forming ativity was assayed immediately after eah purifiation step, and ative material was stored at -20'C without delay. Aliquots of ative frations were analyzed by NaDodSO4/ PAGE. Frations were pooled aording to ativity and homogeneity of the banding pattern and applied to the next olumn. Protein onentration was determined aording to Bradford (13) using bovine serum albumin as standard. The trophozoite extrat was thawed, entrifuged at 150,000 x g for 40 min at 40C, and passed in three bathes of 8 ml ontaining 290 mg of protein eah over a Superdex 75-pg olumn using 20 mm Tris buffer, ph 7.8. Frations ontaining the majority of pore-forming ativity were pooled and applied in one bath of 44 ml ontaining 10 mg of protein to a Mono Q anion-exhange olumn. The olumn was equilibrated with 20 mm Tris buffer, ph 7.8. Adsorbed protein was eluted by washing with the same buffer (40 ml) and by use of NaCl onentration gradients from 0 to 100 mm (160 ml) and from 100 to 300 mm (30 ml) and final washes with 300 mm NaCI (30 ml) and 1 M NaCl (40 ml). Frations with highest pore-forming ativity were pooled, and an aliquot (8 ml, 0.24 mg of protein) was subjeted to a high-performane hydroxylapatite olumn equilibrated with starting buffer (5 mm sodium phosphate, ph 5.8). The olumn was washed with the same buffer and developed with a 100-ml gradient of mm sodium phosphate, ph 5.8. Finally the olumn was washed with 0.5 M sodium phosphate (20 ml). Amino Aid Sequening. Material purified by hromatography on Mono Q was dialyzed against distilled water, onentrated 10-fold by a speed vauum onentrator, and subjeted to reverse-phase HPLC on an Aquapore RP 300 olumn (2.1 x 30 mm; Brownlee) onneted with a 130 A separation system (Applied Biosystems). Elution was done with a linear gradient of 0-70% aetonitrile in water aidified with 0.1% trifluoroaeti aid; the flow rate was 0.2 ml/min, the effluent was monitored by absorbane at 214 nm, and 0.2 ml-frations were olleted. The peak polypeptide fration was applied to a gas-phase protein sequener (model 437 A; Applied Biosystems). Additionally, the pore-forming peptide purified with hydroxylapatite was applied diretly to protein sequening or after gel eletrophoresis and semidry eletroblotting onto a poly(vinylidene difluoride) membrane (Problott; Applied Biosystems). The method did not allow the identifiation of ysteine. Rabbit Antiserum. Antiserum to amoeba peptide was obtained by immunizing a rabbit s.. with 30,ug of the antigen purified by hromatography on Mono Q and emulsified in omplete Freund's adjuvant. Four boosters followed at intervals of 7-10 days, and the animal was bled 1 week after the final injetion. Serum was also olleted before immunization (preimmune serum) and used as ontrol. Gel Eletrophoresis and Blotting. NaDodSO4/PAGE was performed aording to Laemmli (14) with a 4-20% gradient gel and a 4% staking gel. For moleular weight determination, the Triine-NaDodSO4/PAGE for separating small polypeptides, essentially desribed by Shaigger and von Jagow (15), was used. Samples were inubated for 30 min at 20 C in the presene of 1% NaDodSO4; for redution, dithiothreitol was added to a final onentration of 25 mm, and samples were boiled. Redued samples were alkylated with iodoaetamide (50 mm final onentration) for 30 min at 20TC. For glutaraldehyde treatment, samples were exposed to 1% glutaraldehyde for 2 min at 20TC before inubation with NaDodSO4. Gels were stained with silver nitrate. Polypeptides separated by the Triine-NaDodSO4/PAGE were fixed with 5% glutaraldehyde in 0;4 M borate buffer, ph 8.0, for 1 hr before staining. Immunoblots were arried out by use of the semidry blotting tehnique (Multiphor II; Pharmaia LKB) following the instrutions of the manufaturer. Briefly, proteins separated by NaDodSO4/PAGE were eletrotransferred to 0.2 k.m nitroellulose membranes (Shleiher & Shull) followed by bloking with 5% milk powder in phosphate-buffered saline (PBS) for 30 min at 20'C and overnight inubation at 40C with antibodies at a 1:100 dilution in 2.5% milk powder/pbs. After several washes in PBS, the immunoblot was inubated with an anti-rabbit immunoglobulin antibody-peroxidase onjugate (DAKO, Glostrup, Denmark) at 1:400 dilution, washed in PBS, and developed using 4-hloro-1-naphthol as substrate in the presene of H202. RESULTS Charaterization of Pore-Forming Ativity of Pathogeni E. histolytia. Preliminary studies with freeze-thaw lysates of trophozoites onfirmed previous reports on the presene of pore-forming ativity. The 150,000 x g supernatant of the lysate, in the following termed trophozoite extrat, aused depolarization of a valinomyin-indued membrane potential of liposomes. The liposome assay deteted pore-forming ativity in a ph range of , the highest ativity being measured at ph 5.2. To determine whether pore formation depended on the hemial struture of the lipid, trophozoite extrat was preinubated with vesiles of different phospholipids and subsequently tested for ativity in the standard assay. Adsorption of pore-forming material appeared to be affeted by the net harge of the vesile surfae (Fig. 1): All vesiles omposed of phospholipids with negative harge absorbed the ativity, whereas phospholipids without a net harge were ineffetive. Upon separation of trophozoite extrat by moleular sieve hromatography on Superdex 200 pg, a single peak of poreforming ativity was deteted. Ative material was eluted at a position of 14 kda regardless ofwhether the buffer ontained 150 mm NaCl or no salt (Fig. 2). This finding was onfirmed on Superose 12. Chromatography on Superdex 75 pg, whih predominantly separates low-moleular-mass material, revealed the Mr of pore-forming material to be -11 kda. Isolation of the Pore-Forming Peptide from Pathogeni E. histolytia. The starting material was the extrat of 2 x 109 ultured amoebae of the pathogeni strain HM-1:IMSS. An initial moleular sieve hromatography of the extrat on a,5b 100-PIS PE PG S 80 P PM '0 60- ~ < Phospholipid, pg/ml FIG. 1. Absorption of pore-forming ativity from trophozoite extrat by phospholipids. Trophozoite extrat (0.2 mg of protein) was inubated for 3 hr at 40C with vesiles omposed ofphosphatidylinositol (PI), phosphatidylglyerol (PG), phosphatidylserine (PS), sphingomyelin (SM), phosphatidylethanolamine (PE), or phosphatidylholine (PC). As a ontrol trophozoite extrat was inubated without phospholipids. After the inubation period the remaining pore-forming ativity was determined by measuring depolarization of liposome membrane potential. Open symbols indiate negatively harged phospholipids, whereas losed symbols represent the absene of a net harge.

3 Medial Sienes: Leippe et al. Pro. Natl. Aad. Si. USA 88 (1991) 7661 CY OD0 o M Lo N X oi I t CYE I It CY E o OD Coj ai) 0 u < Fration number LO rx' 4. 3 m 2$ 1.2 0< FIG. 2. Moleular sieve hromatography on a Superdex 200 pg olumn of a trophozoite extrat (400 mg of protein) in 20 mm Tris, ph 7.8. A typial elution profile of six separations is shown. Frations (2.5 ml) ontaining pore-forming ativity were identified by the membrane depolarization test and pooled as indiated by the bar. Arrows denote protein standards with moleular mass in kda. Pore-forming ativity was eluted at a position orresponding to 13.7 kda. Superdex 75-pg olumn separated high- and low-moleularmass material from the ative omponent. Further purifiation was ahieved by Mono Q anion-exhange hromatography. The ative material was only weakly retained by the olumn, and the main part of it was eluted with the starting buffer. Pooled frations ontained primarily a polypeptide of -4-5 kda deteted by NaDodSO4/PAGE (Fig. 3). To asertain that this polypeptide represents the pore-forming material and to remove minor ontaminants, an aliquot of this material was hromatographed on a hydroxylapatite olumn. Most pore-forming ativity was eluted at 250 mm sodium phosphate. Again a distint protein band of -4-5 kda was detetable in this fration by NaDodSO4/PAGE (Fig. 3). The isolation to apparent homogeneity ofthe amoeba peptide with this three-step purifiation sheme resulted in a 120-fold purifiation and a reovery of '3% (Table 1). The moleular mass of the peptide was estimated to be 4.7 kda upon Triine-NaDodSO4/PAGE under nonreduing onditions and after redution and alkylation (data not shown). Partial Primary Struture. The amoeba peptide obtained from the Mono Q olumn was purified further either by hydroxylapatite hromatography or by HPLC and subjeted to protein sequening. The NH2-terminal sequene of both samples was idential and revealed a primary struture similar to that of melittin from Apis florea (Fig. 4). Reversible Oligomerization of Amoeba Peptide. To examine its moleular organization after insertion into a membrane, amoeba peptide adsorbed to phosphatidylinositol vesiles was subjeted to NaDodSO4/PAGE. Without hemial rosslinking, the peptide exhibited monomeri behavior. However, treatment of the liposome-bound peptide with glutaraldehyde for 2 min revealed the presene of homooligomers that were deteted by immunoblot analysis by using an antiserum to the amoeba peptide (Fig. 5). In addition to the monomer, higher-moleular-weight entities were seen that appeared to be dimers up to hexamers. Amoeba peptide kda SWWV mp A B C D FIG. 3. Amoeba peptide at different stages of purifiation. Aliquots of pore-forming material were analyzed after eah purifiation step by NaDodSO4/PAGE on a 4-20% gradient gel under nonreduing onditions (silver stain). Lanes: A, trophozoite extrat (37 Mug); B, Superdex 75-pg fration (2.3 Mg); C, Mono Q fration (0.3,ug); D, hydroxylapatite fration (0.1 Mig). Standards are indiated in kda. in solution showed a similar behavior, exept that oligomerization was less extensive. Treatment of a solution of ovalbumin with glutaraldehyde under the same onditions did not produe rosslinked oligomers. DISCUSSION In the present study we isolated a pore-forming peptide from pathogeni E. histolytia. The amoeba peptide desribed here is probably idential to amoebapore, the pore-forming protein partially purified by Rosenberg et al. (8), beause in both purifiation proedures the starting material was the same amoeba strain, the first two purifiation steps were nearly idential, and the ativity assay was very similar. However, the moleular mass of the pore-forming omponent estimated in this study differs from that reported (8). These authors proposed that in the native state amoebapore onstitutes a dimer of kda, whereas in our study moleular sieve hromatography on different matries yielded a moleular mass of kda for the ative peptide. NaDodSO4/PAGE results have led to the suggestion (8) that the moleular mass of the monomer was 14 kda, although the authors noted that the partial purifiation and small amounts of material preluded doumentation of purity on a polyarylamide gel. In ontrast, we found the pore-forming peptide to have a moleular mass of 4.7 kda upon NaDod- S04/PAGE. These disparate results may be from the use of different matries for moleular sieve hromatography and from the insuffiient resolving power of standard Na- DodSO4/PAGE for peptides <15 kda (15). Our moleular sizing experiments do not allow us to deide whether the amoeba peptide in its native state ours as a dimer or whether the moleule in its monomeri form possesses an unusually high Stokes' radius. A similar phenomenon was Table 1. Purifiation of the amoeba peptide Total protein, Total ativity,* Speifi ativity, Material mg units x 10-3 units x 10-3 per mg Yield, % Purifiation Trophozoite extrat , Superdex 75 pg 10 6, Mono Q 0.7 2,080 2, Hydroxylapatitet , *Data were estimated by measuring liposome depolarization. tdata were extrapolated from an aliquot.

4 7662 Medial Sienes: Leippe et al. Pro. Nad. Aad. Si. USA 88 (1991) Amoeba peptide G E I L X N LXTG L I NTLE N L L T X K G A D Amoeba peptide Melittin Melittin GIGAILKVLATGLP TLISWIS KN1KRKQ* FIG. 4. NH2-terminal amino aid sequene of the amoeba peptide ompared with the sequene of melittin of Apis florea (one-letter notation of amino aids is used). Idential residues are indiated by vertial bars. When identifiation was ambiguous or impossible, the residue is represented as X. The asterisk indiates an amidated arboxyl terminus. For optimal alignment a gap was introdued into the sequene of melittin. In the overlap of 25 amino aids, 40%o of the residues are idential. desribed for the 4-kDa antibaterial peptide eropin A from silk moths, whih eluted from a alibrated sizing olumn at a position orresponding to -14 kda (16). However, our rosslinking experiments with glutaraldehyde suggest that indeed the amoeba pore-forming peptide tends to undergo self-assoiation and to form reversible oligomers. The amoeba peptide expresses no speifiity for a ertain lipid struture, but it preferentially inserts into negatively harged phospholipid bilayers irrespetive of the nature of the headgroups. Suh a behavior is known from a variety of ytolysins and was proposed to be based on a ommon ationi site that determines the lyti ativity by enhaning the interation with negatively harged vesiles or ell surfaes (17). Examples of these kinds ofpeptides are sarotoxin IA, an antibaterial peptide of flesh flies (18), and melittin, the membranolyti polypeptide of the venoms from bee speies (19, 20). The partial primary struture of the amoeba peptide determined by protein sequening shows a remarkable degree of similarity to melittin, espeially to that of Apisflorea (Fig. 4). The primary struture of this melittin differs from that of the ommon honeybee, Apis mellifera, at five positions (21). Projetion of the NH2-terminal amino aids 1-20 of the amoeba peptide as an a-helial wheel (22) showed a segregation of hydrophili and hydrophobi amino aids similar to melittin (Fig. 6), suggesting similarity in seondary struture. The amphiphili a-helial struture is known for many mem- kda Protein stain immunoblot "O g o ont. +lip. +lip. +GA ant. --+lp. +Iip. +GA FIG. 5. Reversible oligomerization of the amoeba peptide bound to liposomes. Amoeba peptide purified by Mono Q hromatography was subjeted diretly (ont.) or after adsorption to phosphatidylinositol vesiles (+ lip.) to NaDodSO4/PAGE on a 4-20% gradient gel under nonreduing onditions (0.24,ug of peptide per lane, silver stain). Inubation with 1% glutaraldehyde (+ GA) for 2 min at 20 C reveals rosslinked oligomers of the peptide. The orresponding immunoblot was developed with an antiserum to the amoeba peptide. Glutaraldehyde treatment appears to inrease staining intensity of the peptides. Standards are indiated in kda at left P G 21 ( ) KNKRKQ* FIG. 6. Projetion as a-helial wheel of the NHrterminal sequene (residues 1-20) of the amoeba peptide and of the sequene of melittin from Apis florea (single-letter ode is used). The helix is viewed from one end with the perimeter of a wheel orresponding to the bakbone of the peptide. Residues are spaed 1000 of ar apart to give 3.6 residues per turn. Amino aids with hydrophobi side hains are irled. The asterisk indiates an amidated arboxyl terminus. branolyti peptides and is thought essential for their ativity: beside the melittins (23), ytolysins like 8-hemolysin from Staphyloous aureus (24), mastoparan from the venom of wasps, and the bombolitins from the venom of bumblebees (25), antibaterial peptides like eropins (26) and sarotoxins (27) from inset immune hemolymph, and magainins from frog ventral skin (28) all have this ommon strutural feature. For these small peptides a similar mode of ation has been proposed that results in the formation of ion-hannel pores spanning the membrane (29, 30). Suh helial peptides should be at least 20 residues in length to allow a perturbation of the lipid bilayer (31). Its amphiphili nature would render part of the outside of the a-helix hydrophobi and apable of interating with the hydroarbon hains of the lipid bilayer. The hydrophili part of the inserted moleule would ause reorientation of the ordered bilayer. A lipid membrane so modified no longer onstitutes a permeability barrier but allows transmembrane ion flow to our. The hemially different peptides with membranolyti ativity mentioned above and studies of syntheti amphiphili a-helial peptides (31, 32) suggest that the seondary struture, rather than the primary struture, is essential for their ytolyti funtion. The ytolyti ativity of most pore-forming proteins desribed so far is based on the nonovalent aggregation of these moleules in the target membrane (4). The ninth omponent of omplement (33, 34) and perforin (2, 35, 36), the pore-forming protein of killer lymphoytes, are wellharaterized members of this group. Both proteins form large, tubular strutures by polymerization of protomers within the lipid bilayer. Models have been proposed aording to whih a similar ytolyti mehanism applies also to smaller peptides (31, 37). With respet to the pore-forming ativity of E. histolytia, the data obtained using planar bilayers (7-9, 38) suggest that an aggregation of protomers ourred during pore formation. Oligomers might already be present in aqueous solution (38). Keller et al. (9) postulated that amoebapore is a member of the barrel-stave lass of pore formers like alamethiin and melittin in that pores onsist of omplexes of variable numbers of protomers. The observation that the amoeba peptide tends to self-assoiate in solution and when inorporated into phospholipid bilayers may onstitute diret experimental evidene for this hypothesis. In addition to Entamoeba histolytia, the two protozoan parasites Naegleria fowleri (39) and Trypanosoma ruzi (40)

5 Medial Sienes: Leippe et al. also possess pore-forming polypeptides. Whereas T. ruzi appears to use the pore-forming ativity for evasion from the phagosome into the ytoplasm of the host ell, the poreforming peptide of E. histolytia is thought to mediate the ontat-mediated killing of host ells by pathogeni amoebae. Note Added in Proof. Moleular loning of DNA and protein sequening of enzymatially produed peptide fragments reveal the moleular mass of amoeba peptide to be onsiderably larger (8 kda) than indiated by its eletrophoreti behavior in NaDodSO4/PAGE. We thank Jurgen Sievertsen for tehnial assistane. 1. Ravdin, J. I. (1989) Pathol. Immunopathol. Res. 8, Tshopp, J. & Nabholz, M. (1990) Annu. Rev. Immunol. 8, Henkart, P. A. (1985) Annu. Rev. Immunol. 3, Bhakdi, S. & Tranum-Jensen, J. (1987) Rev. Physiol. Biohem. Pharmaol. 107, Ojius, D. M. & Young, J. D.-E. (1990) Parasitol. Today 6, Lynh, E. C., Rosenberg, I. M. & Gitler, C. (1982) EMBO J. 1, Young, J. D.-E., Young, T. M., Lu, L. P., Unkeless, J. C. & Cohn, Z. A. (1982) J. Exp. Med. 156, Rosenberg, I., Bah, D., Loew, L. M. & Gitler, C. (1989) Mol. Biohem. Parasitol. 33, Keller, F., Hanke, W., Trissl, D. & Bakker-Grunwald, T. (1989) Biohim. Biophys. Ata 982, Diamond, L. S., Harlow, D. R. & Cunnik, C. C. (1978) Trans. R. So. Trop. Med. Hyg. 72, Loew, L. M., Rosenberg, I., Bridge, M. & Gitler, C. (1983) Biohemistry 22, Pik, U. (1981) Arh. Biohem. Biophys. 212, Bradford, M. M. (1976) Anal. Biohem. 72, Laemmli, U. K. (1970) Nature (London) 227, Shagger, H. & von Jagow, G. (1987) Anal. Biohem. 166, Steiner, H., Hultmark, D., Engstr6m, A., Bennih, H. & Boman, H. G. (1981) Nature (London) 292, Kini, R. M. & Evans, H. J. (1989) Int. J. Peptide Protein Res. 34, Pro. Natl. Aad. Si. USA 88 (1991) Nakajima, Y., Qu, X. & Natori, S. (1987) J. Biol. Chem. 262, Beshiashvili, G. & Seelig, J. (1990) Biohemistry 29, Stankowski, S. & Shwarz, G. (1990) Biohim. Biophys. Ata 1025, Kreil, G. (1973) FEBS Lett. 33, Shiffer, M. & Edmundson, A. B. (1%7) Biophys. J. 7, Dempsey, C. E. (1990) Biohim. Biophys. Ata 1031, Mellor, I. R., Thomas, D. H. & Sansom, M. S. P. (1988) Biohim. Biophys. Ata 942, Argiolas, A. & Pisano, J. J. (1985) J. Biol. Chem. 260, Fink, J., Boman, A., Boman, H. G. & Merrifield, R. B. (1989) Int. J. Peptide Protein Res. 33, Okada, M. & Natori, S. (1985) Biohem. J. 229, Marion, D., Zasloff, M. & Bax, A. (1988) FEBS Lett. 227, Wade, D., Boman, A., Wahlin, B., Drain, C. M., Andreu, D., Boman, H. G. & Merrifield, R. B. (1990) Pro. Nati. Aad. Si. USA 87, Berkowitz, B. A., Bevins, C. L. & Zasloff, M. A. (1990) Biohem. Pharmaol. 39, Lear, J. D., Wasserman, Z. R. & DeGrado, W. F. (1988) Siene 240, Parente, R. A., Nadasni, L., Subbarao, N. K. & Szoka, F. C., Jr. (1990) Biohemistry 29, Podak, E. R., Tshopp, J. & Muller-Eberhard, H. J. (1982) J. Exp. Med. 156, Tshopp, J., Muller-Eberhard, H. J. & Podak, E. R. (1982) Nature (London) 298, Podak, E. R. (1986) J. Cell. Biohem. 30, Tshopp, J. (1984) J. Biol. Chem. 259, Raghunathan, G., Seetharamulu, P., Brooks, B. R. & Guy, H. R. (1990) Proteins 8, Young, J. D.-E. & Cohn, Z. A. (1985) J. Cell. Biohem. 29, Young, J. D.-E. & Lowrey, D. M. (1989) J. Biol. Chem. 264, Andrews, N. W., Abrams, C. K., Slatin, S. L. & Griffiths, G. (1990) Cell 61,