The Neutral Cysteine Proteinase of Entamoeba histolytica Degrades ...

508

The Neutral Cysteine Proteinase of Entamoeba histolytica Degrades IgG and Prevents Its Binding Vinh Q. Tran,* D. Scott Herdman, Bruce E. Torian, and Sharon L. Reed

Departments of Pathology and Medicine, University of California at San Diego, San Diego, California

Patients infected with Entamoeba histolytica generate specific IgG that does not prevent invasive amebiasis or recurrent infection. Studies investigated whether the effectiveness of the human humoral response was limited by cleavage of IgG by the extracellular neutral cysteine proteinase of E. histolytica trophozoites, one of the first amebic products to interact directly with components of host defenses. Purified proteinase cleaved polyclonal human and monoclonal murine IgG in a dosedependent manner. Peptide sequencing of the major cleavage fragment(s), which contained the protein A binding site, suggested that cleavage occurred near the hinge region. Intact trophozoites also cleave IgG in both growth media and serum-free media. Cleaved monoclonal antibody to a 29kDa surface antigen of E. histolytica bound to trophozoites 83.5% { 6.7% less than did uncleaved antibody. These results suggest that cleavage of IgG by the extracellular cysteine proteinase may limit the effectiveness of the host humoral response.

Invasive amebiasis continues to be a major public health problem in developing countries, affecting ú50 million people yearly. Although up to 10% of the world’s population is infected with Entamoeba species, õ1% actually develop invasive disease. The large discrepancy between the number of infected patients and those with invasive amebiasis can be explained in part by the recent findings that there are two morphologically identical species: Entamoeba histolytica, which is capable of invasion, and Entamoeba dispar, which is not [1]. One virulence factor that may enable E. histolytica to invade is an extracellular neutral cysteine proteinase that degrades extracellular matrix and complement components [2, 3]. Cysteine proteinases in amebic lysates degrade IgA [4] and thus might also affect the efficacy of the humoral response by cleavage of IgG. Specific IgG responses develop in ú95% of patients with invasive amebiasis or even asymptomatic colonization with E. histolytica [5]. Despite this response, invasive amebiasis follows intestinal colonization, and recurrent amebic liver abscess, although rare, does occur. To determine if the systemic host immune response might be limited by proteolytic degradation of IgG, we investigated the interactions of polyclonal and monoclonal IgG with the purified neutral cysteine proteinase and live trophozoites.

Received 27 June 1997; revised 10 September 1997. Presented in part: 13th Seminar on Amebiasis, 29 – 31 January 1997, Mexico City (abstract: Arch Med Res 1997; 28:178 – 9). Grant support: NIH (AI-28035 and DK-35108 to S.L.R.); Lucille P. Markey Charitable Trust (to S.L.R.). Reprints or correspondence: Dr. Sharon Reed, Division of Infectious Diseases, UCSD Medical Center, 200 W. Arbor Dr., San Diego, CA 92103-8416 ([email protected]). * Present affiliation: Ross University School of Medicine, Portsmouth, West Indies. The Journal of Infectious Diseases 1998;177:508–11 q 1998 by The University of Chicago. All rights reserved. 0022–1899/98/7702–0040$02.00

/ 9d3f$$fe26

12-23-97 08:53:41

jinfa

Materials and Methods E. histolytica trophozoites. Axenic E. histolytica (strain HM1:IMSS; American Type Culture Collection, Rockville, MD) was grown in TYI-S-33 medium containing 15% bovine serum [1] and subcultured twice weekly. Amebae were purified by chilling the flasks, centrifuging the trophozoites at 300 g for 10 min, and washing three times in PBS. Immunoglobulins. Human polyclonal IgG (Sigma, St. Louis) and monoclonal anti–bovine serum albumin (Sigma) were suspended at 1 mg/mL in PBS. Monoclonal antibody FP31 (IgG2a) was generated against a glutathione-S-transferase fusion protein of the 29-kDa thiol-rich surface antigen of E. histolytica as described [6]. Immunoglobulins were radiolabeled with 125I to a specific activity of 106 cpm/mg of protein by the iodogen method [3]. Proteinase purification. Proteinase was purified from amebic trophozoites or secretions as described [2], and purity was confirmed on 10% SDS-polyacrylamide gels copolymerized with 0.1% gelatin [2] and silver-stained gels. The proteinase activity was determined by cleavage of the synthetic peptide, Boc-argininearginine-4-amino-7-methylcoumarin (Z-Arg-Arg-AMC; Enzyme System Products, Dublin, CA), as described previously [2, 3]. The activity of the proteinase was measured before each experiment. Proteolytic digestion of IgG. Aliquots of unlabeled polyclonal IgG (2 mg) or radiolabeled murine monoclonal IgG (106 cpm added to 1 mg of cold IgG) were incubated for 1–18 h at 377C with 0– 50 U of purified proteinase/mg of IgG. Samples were boiled in the presence of the 10 mM specific inhibitor, P35017 (gift of Prototek, Dublin, CA), and reduced by adding an equal volume of 0.1 M Tris, 2.5% b-mercaptoethanol, and 2% SDS. Controls included samples in which the proteinase was inhibited with 10 mM P35017 for 30 min at room temperature before addition of IgG. Samples were electrophoresed by SDS-PAGE (15% or 7.5%– 15% gradient gels) under reducing conditions. Gels were dried and autoradiographed at 0707C on Kodak XAR-5 film. For immunoblots, gels were transferred to nitrocellulose, and membranes were blocked for 1 h with Tris-buffered saline (TBS) containing 1% nonfat milk and 0.5% Tween 20, washed twice in TBS/Tween, and incubated for 1.5 h with a 1:1000 dilution of alkaline phospha-

UC: J Infect

JID 1998;177 (February)

Concise Communications

509

added. Controls included an unrelated monoclonal antibody to bovine serum albumin (Sigma).

Results

Figure 1. Dose-dependent cleavage of polyclonal and monoclonal IgG. A, Iodinated monoclonal antibody FP31 (106 cpm) was added to 1 mg of cold IgG and incubated with 0 – 50 U of proteinase/mg of IgG for 18 h at 377C; cleavage products were visualized on autoradiographs of reduced polyacrylamide gels. Dose-dependent production of cleavage fragment (F) from heavy chain (HC) is seen. Light chain (LC) remains uncleaved. B, Western blot of proteinase-cleaved polyclonal IgG developed with protein A. Human polyclonal IgG (2.5 mg) was incubated for 18 h with 5 – 40 U of proteinase/mg of IgG. Resulting polyacrylamide gel was transferred to nitrocellulose and developed with alkaline phosphatase – labeled protein A. P Å proteinase alone. HC and F are shown.

tase–conjugated protein A (Zymed, South San Francisco, CA). After washing, bands were detected with nitroblue tetrazolium/5bromo-4-chloro-3-indolyl phosphate. Peptide sequencing. Human polyclonal IgG (15 mg) was cleaved overnight with 600 U of purified proteinase, electrophoresed under reducing conditions, and transferred to 0.45-mm polyvinylidenefluoride membranes (Immobilon; Millipore, Bedford, MA). Bands were then detected by Coomassie blue staining, excised, and sequenced by Edman degradation (491 Procise sequencer; Applied Biosystems, Foster City, CA). Binding of IgG to live trophozoites. Radiolabeled monoclonal antibody FP31 (106 cpm) was added to amebic cultures in TYI-S33 overnight or to MEM-CH (Eagle MEM, HEPES buffer solution, ascorbic acid, and cysteine [3]) in 3-mL vials for 4 h at 377C. Viability was assessed with trypan blue staining in duplicate samples in MEM-CH alone. Amebae were pelleted at 300 g for 10 min and washed three times with PBS. Aliquots of the pellets and media were prepared for electrophoresis as detailed above. The effect of proteolytic cleavage on binding of IgG to live amebic trophozoites was assessed by incubating 125I-labeled monoclonal antibody FP31 (1.34 mg) for 18 h with purified proteinase (50 U/mg of antibody). At least 60% cleavage of the heavy chain was confirmed by scanning the autoradiograph with an Apple One scanner (Apple Computers, San Jose, CA) and Image software (Image 1.59 freeware; NIH, Bethesda, MD). The cleaved and uncleaved monoclonal antibodies were subsequently incubated with live trophozoites (2 1 105 cells/sample) in triplicate for 0–2 h in MEM-CH at 377C. The cells were washed three times with PBS to remove excess unbound label, and the washes were pooled for counting. The percentage of binding was calculated as the counts per minute bound to the washed pellet divided by the total counts

/ 9d3f$$fe26

12-23-97 08:53:41

jinfa

Dose-dependent cleavage of IgG. After incubation of polyclonal or monoclonal antibody with 10 – 50 U of proteinase/mg of protein, a major cleavage product of 30 – 35 kDa was detected (figure 1A). Cleavage was completely inhibited by preincubation with the specific cysteine proteinase inhibitor P35017 (data not shown). Similar results were obtained with polyclonal human IgG (figure 1B) and IgG1 and IgG2 subclasses of murine monoclonal antibodies (data not shown). Localization of the cleavage site of IgG. To identify the specific site of cleavage, degradation products of human polyclonal IgG were probed with antibodies to protein A. After degradation of polyclonal IgG with 5 – 40 U of proteinase/mg of protein, Western blots of the cleavage products developed with alkaline phosphatase – conjugated protein A showed a single degradation product (figure 1B). When the cleavage product was transferred to polyvinylidenefluoride membranes for microsequencing, however, multiple sequences were obtained, suggesting that cleavage of the heavy chain occurred at more than one site. The major fragment contained the sequence XPro-Pro-X-Pro-Ala-Pro, which is located in the hinge region. Cleavage of IgG by live trophozoites. To ensure that live trophozoites also cleaved IgG, 125I-labeled monoclonal antibody to the 29-kDa surface antigen was added to washed trophozoites in serum-free MEM-CH, in which ú90% viability is maintained for 6 h [3], or amebic cultures in TYI-S-33. Autoradiographs of the resulting supernatants revealed a cleavage product of 30 – 35 kDa, similar to the cleavage product produced by purified proteinase (figure 2). Almost complete degradation of the heavy chain of IgG was detected in amebic pellets. Equivalent results were detected in amebic growth media (TYI-S-33), although incubation for at least 24 h was required (data not shown). Binding of proteinase-cleaved IgG. To determine if proteinase-cleaved IgG could still bind to the amebic surface, we incubated purified amebae (2 1 105) in triplicate with cleaved

Figure 2. Cleavage of iodinated monoclonal antibody FP31 added to intact amebic trophozoites. After 4 h of incubation in serum-free Eagle MEM, HEPES buffer solution, ascorbic acid, and cysteine media, trophozoites were pelleted and extensively washed, and IgG fragments were detected on autoradiographs of pellets (P) and supernatants (S). C Å control of antibody incubated in media alone. Similar cleavage fragment (F) from heavy chain (HC) is seen in supernatant fractions. LC Å light chain.

UC: J Infect

510


or uncleaved radioactive monoclonal antibodies to the 29-kDa antigen. Binding by radiolabeled monoclonal antibody to bovine serum albumin was compared as a control. By 30 min, the cleaved antibody bound 83.5% { 6.7% less to trophozoites than did uncleaved antibody (P õ .0001). At 30 min, only 2.5% of added anti – bovine serum albumin was bound to the amebic pellet, likely representing nonspecific trapping. Discussion Invasion or even colonization by E. histolytica trophozoites stimulates both systemic and local antibody responses, in contrast to colonization with E. dispar [5]. Clearance of luminal parasites or protection afforded by systemic IgG has been difficult to demonstrate, however, and the level of antibody response correlates with the length of disease, not with the severity or clinical response to infection [5]. Similarly, protection of hamsters against liver abscesses does not correlate with the serum antibody response, although SCID mice were passively protected against infection by intraperitoneal injection of rabbit polyclonal immune serum [7]. The effect of passive antibody in the SCID mouse model may reflect both the quantity of antibody and the direct delivery of a bolus of antibody at the site of inoculation of the trophozoites. The primary purpose of the amebic cysteine proteinase is most likely to break down endocytosed proteins. Significant quantities of cysteine proteinases are also released extracellularly [2] and thus are likely to be the first amebic products to interact with components of the host defense. The neutral cysteine proteinase has been shown to play a role in amebic invasion through degradation of components of the extracellular matrix, including collagen and elastin [2], activation of complement components by cleavage of C3 [3], and degradation of the anaphylatoxins, C3a and C5a [3]. Kelsall and Ravdin [4] concluded that cysteine proteinases were responsible for the degradation of IgA by amebic lysates on the basis of the effect of specific inhibitors. Cysteine proteinases are also a major virulence factor that differentiates E. histolytica from E. dispar. Clinical isolates of E. histolytica release significantly more cysteine proteinase activity and contain unique cysteine proteinase gene(s) [8]. If the cysteine proteinases of E. histolytica degrade IgG, this property may help explain the ineffectiveness of humoral defenses against amebiasis. In the current studies, we showed that the purified amebic cysteine proteinase cleaved the heavy chain of both human polyclonal and murine monoclonal antibody in a dose-dependent manner (figure 1A). The major cleavage fragment of 30 – 35 kDa contained the protein A binding site (figure 1B). Microsequencing of the N-terminal fragment demonstrated, however, that several size-related fragments were produced. On the basis of the proline-rich sequence of one fragment, a region near the hinge was the major cleavage site. To determine whether live amebic trophozoites could also cleave IgG, we investigated the cleavage of a radiolabeled

/ 9d3f$$fe26

12-23-97 08:53:41

jinfa


monoclonal antibody to the 29-kDa thiol-rich surface antigen [6]. Trophozoites incubated with IgG in the serum-free media, MEM-CH [3], completely degraded the heavy chain in both supernatant and pellet fractions. The intensity of the light chain band was slightly decreased in the pellet fraction, most likely the result of internalization and subsequent degradation of bound immunoglobulin fragments by whole amebae [9]. Proteinase cleavage also prevented binding of a monoclonal antibody to an amebic surface antigen. The monoclonal antibody (FP31) against the 29-kDa surface antigen was precleaved with 50 U of proteinase/mg of protein, resulting in ú60% cleavage of the heavy chain. After 30 min of incubation with live trophozoites, binding of the cleaved antibody was reduced by 89% compared with intact antibody and was equivalent to the nonspecific binding of a monoclonal antibody to bovine serum albumin. Although capping and release of polyclonal antibody by E. histolytica trophozoites has been demonstrated [9], this phenomenon would not explain the significant difference in binding of cleaved and uncleaved IgG. A number of potential human pathogens degrade immunoglobulins. The proteolytic digestion of IgA1 by mucosally invasive bacteria has been the best characterized [10]. These extracellular enzymes specifically cleave the heavy-chain hinge region of IgA1, and their presence is closely linked to pathogenicity of a number of bacteria [10]. Cysteine proteinases in amebic lysates have also been shown to cleave IgA1 [4]. Both protozoa and helminths have been shown to release proteinases that cleave IgG. Trypanosoma cruzi trypomastigotes bind nonimmune IgG through the Fab fragment and cleave the Fc fragment by the action of the major cysteine proteinase, cruzipain [11]. Schistosoma mansoni binds immunoglobulins by the Fc receptor and degrades the Fab portion into small peptides by the action of at least two proteinases [12]. Cysteine proteinases also play a role in the degradation of IgG by Fasciola hepatica [13], Trichomonas vaginalis [14], and Tritrichomonas foetus [15]. These data demonstrate that E. histolytica trophozoites are capable of cleaving bound and fluid-phase IgG. Cleavage of monoclonal antibodies specific to major surface antigens prevents their binding. Although in vivo studies would be required for definitive proof that cleavage of IgG by the amebic proteinase is a key mechanism of immune evasion, release of extracellular proteinases that cleave IgG is a common property of invasive parasites and may contribute in part to the failure of antibody alone to protect against amebic invasion. Acknowledgments

We thank Charles Davis and Lynette Corbeil for helpful discussions. References 1. Clark CB, Diamond LS. Ribosomal RNA genes of ‘‘pathogenic’’ and ‘‘nonpathogenic’’ Entamoeba histolytica are distinct. Mol Biochem Parasitol 1991; 49:297 – 302.

UC: J Infect



2. Keene WE, Pettit MG, Allen S, McKerrow JH. The major neutral proteinase of Entamoeba histolytica. J Exp Med 1986; 163:536 – 49. 3. Reed SL, Ember JA, Herdman DS, DiScipio RG, Hugli TE, Gigli I. The extracellular neutral cysteine proteinase of Entamoeba histolytica degrades anaphylatoxins C3a and C5a. J Immunol 1995; 155:266 – 74. 4. Kelsall BL, Ravdin JI. Degradation of human immunoglobulin A by Entamoeba histolytica. J Infect Dis 1993; 168:1319 – 22. 5. Jackson TFHG, Gathiram V, Simjee AE. Seroepidemiological study of antibody responses to the zymodemes of Entamoeba histolytica. Lancet 1985; 2:716 – 9. 6. Reed SL, Flores BM, Batzer MA, et al. Molecular and cellular characterization of the 29-kDa peripheral membrane protein of Entamoeba histolytica: differentiation between pathogenic and nonpathogenic isolates. Infect Immun 1992; 60:542 – 9. 7. Cieslak PR, Virgin HW, Stanley SL. A severe combined immunodeficient (SCID) mouse model for infection with Entamoeba histolytica. J Exp Med 1992; 176:1605 – 9. 8. Reed SL, Bouvier J, Pollack AS, et al. Cloning of a virulence factor of Entamoeba histolytica: pathogenic strains possess a unique cysteine proteinase gene. J Clin Invest 1993; 91:1532 – 40.

511

9. Calderon J, Munoz M, Acosta HM. Surface redistribution and release of antibody-induced caps in Entamoebae. J Exp Med 1980; 151:184 – 93. 10. Plaut AG. The IgA1 proteases of pathogenic bacteria. Annu Rev Microbiol 1983; 37:603 – 22. 11. Bontempi E, Cazzulo JJ. Digestion of human immunoglobulin G by the major cysteine proteinase (cruzipain) from Trypanosoma cruzi. FEMS Microbiol Lett 1990; 70:337 – 42. 12. Auriault C, Ouaissi MA, Torpier G, Eisen H, Capron A. Proteolytic cleavage of IgG bound to the Fc receptor of Schistosoma mansoni schistosomula. Parasite Immunol 1981; 3:33 – 44. 13. Carmona C, Dowd AJ, Smith AM, Dalton JP. Cathepsin L proteinase secreted by Fasciola hepatica in vitro prevents antibody-mediated eosinophil attachment to newly excysted juveniles. Mol Biochem Parasitol 1993; 62:9 – 18. 14. Provenzano D, Alderete JF. Analysis of human immunoglobulin – degrading cysteine proteinases of Trichomonas vaginalis. Infect Immun 1995; 63:3388 – 95. 15. Talbot JA, Nielsen K, Corbeil LB. Cleavage of proteins of reproductive secretions by extracellular proteinases of Tritrichomonas foetus. Can J Microbiol 1991; 37:384 – 90.

Molecular Analysis of Plasmodium vivax Relapses Using the MSP1 Molecule as a Genetic Marker Karin Kirchgatter and Hernando A. del Portillo

Departamento de Parasitologia, Instituto de Cieˆncias Biome´dicas, Universidade de Saõ Paulo, and Superintendeˆncia de Controle de Endemias (SUCEN), Saõ Paulo, Brazil

Plasmodium vivax has hepatocytic dormant stages, hypnozoites, that cause relapses. This work compared paired isolates from primary attacks and relapses obtained from 10 individuals in Brazil using the merozoite surface protein 1 gene, PvMSP1, as a genetic marker. Four samples from primary attacks contained genetically mixed parasites harboring the 2 major PvMSP1 allelic forms. PCR revealed the presence of these 2 forms in the relapse parasites of 2 patients, demonstrating that the activation of hypnozoites is not clonal. DNA sequences from paired primary/relapse samples demonstrated that the parasites from the primary attack are identical to those in relapse samples in which the same allele forms were detected in both infections. Studies on the naturally acquired humoral immune responses of these patients against a recombinant protein expressing the C-terminus PvMSP1 demonstrated an increase in the titers, affinity maturation, and predominance of the IgG1 subclass during the relapse.

Plasmodium vivax is the most widely distributed human malarial parasite, causing Ç35 million cases annually [1]. In some parts of the world, including Brazil, where it reaches 70% of all the yearly malaria cases, this is the most prevalent species.

Received 9 June 1997; revised 5 September 1997. Presented in part: XII annual meeting of the Brazilian Society of Protozoology, Caxambu, Brazil, November 1996 (abstract 047). Informed consent was obtained from all patients involved in this study. DNA sequences from all these samples can be accessed (GenBank nos. AF002181 – AF002190). Grant support: Fundaçaõ de Amparo a` Pesquisa do Estado de Saõ Paulo (94/0227-2) and SUCEN. Reprints or correspondence: Dr. Hernando A. del Portillo, Dept. de Parasitologia, Instituto de Cieˆncias Biome´dicas, Universidade de Saõ Paulo, Av. Lineu Prestes 1374, Saõ Paulo, SP 05508-900, Brazil ([email protected]). The Journal of Infectious Diseases 1998;177:511–5 q 1998 by The University of Chicago. All rights reserved. 0022–1899/98/7702–0041$02.00

/ 9d3f$$fe26

12-23-97 08:53:41

jinfa

P. vivax malaria relapses [2] as clinical attacks that appear after schizonticidal treatment and that are partly responsible for the large socioeconomic burden of this human malaria. Unfortunately, the molecular basis of this phenomenon and its biologic significance remain elusive, mainly because of the difficulties in obtaining parasites from relapse cases. One study has been published to date on the molecular analysis of paired primary and relapse isolates [3]. Significantly, it demonstrated that the strains causing the primary attack were not genetically different from those arising during relapses. In the present study, we compared paired primary and relapse parasites from 10 Brazilian patients by use of a polymorphic segment of the P. vivax merozoite surface protein 1 (PvMSP1) gene as the genetic marker. The naturally acquired IgG immune responses of these patients against a recombinant protein expressing the C-terminal region of PvMSP1 were also determined.

UC: J Infect