J. Eukrrry,r Mirrohiol., 49(6), 2002 pp. 4 5 4 4 5 9 0 2002 by the Society of Protozoologists
An Extracellular monoADP-ribosyl Transferase Activity in Entamoeba histolytica Trophozoites PATRICIA DELGADO-CORONA, GUADALUPE MART~NEZ-CADENA,ANGEL H. ALVAREZ, HORACIO E. TORRES-CALZADA and EVA E. AVILA Instituto de hvestigacidn en Biologia Experimental, Facultad de Quimica, Universidad de Guunajuato, P 0 Box 187, Guanajuato, Gto., Mgxico CP36000
ABSTRACT. Due to the important role of monoADP-ribosyl transferases in physiological and pathological events, we investigated whether the protozoan parasite Entumoebu histolytica had monoADP-ribosyl transferase activity. Reactions were initiated using amebafree medium as the source of both enzyme and ADP-ribosylation substrate(s) and P2P]NAD! as source oL’ ADP-ribose. Proteins were analyzed by electrophoresis, and [32P]-labeledproteins were detected by autoradiography. Using the crude extracellular medium, a major labeled product of Mr 37,000 was observed. The yield of this product was reduced markedly using medium from Brefeldin A-treated trophozoites, indicating that the extracellular monoADP-ribosyl transferase andor its substrate depended on vesicular transport. The labeling of the 37-kDa substrate was dependent on reaction time, temperature, pH, and the ratio of unlabeled NAD+ to [s2P]NAD+. After two purification steps, several new substrates were observed, perhaps due to their enrichment. The reaction measured ADPribosylation since [“C-carbonyl]NAD’ was not incorporated into ameba substrates and a 75-fold molar excess of ADP-ribose caused no detectable inhibition of the monoADP-ribosyl transferase reaction. On the basis of sensitivity to NH,OH, the extracellular monoADPribosyl transferase of E. hisrolytica may be an arginine-specific enzyme. These results demonstrate the existence in E. hisrolyticn of at least one extracellular monoADP-ribosyl transferase, whose localization depends upon a secretion process. Key Words. b.DP-ribosylation, ameba, Brefeldin A, C3 exoenzyme, protozoa, Rho protein.
M
onoADP-RIBOSYLATION is the catalyzed transfer of ADP-ribose from NAD+ to an acceptor protein. It is a post-translational modification implicated in the regulation of key biological functions in eukaryotes and prokaryotes. The reaction can be reversed by the so-called ADP-ribosyl hydrolases (Moss, Jacobson, and Stanley 1985;Moss et al. 1992;Saari et al. 1986). The expression of monoADP-ribosyl transferases (mADPRT) in mammalian tissues has so far been reported mainly in muscle, hematopoietic, and epithelial cells. Many of these enzymes are plasma membrane proteins, some being shed from the cell surface (Moss et al. 1999; Okazaki and Moss 1998). These enzymes presumably participate in signal transduction (EhretHilberer et al. 1992;Lupi, Corda, and Di Girolamo ZOOO), proliferation and differentiation of cultured myoblasts (Kharadia et al. 1992; Yuan et al. 1999), and regulation of T lymphocyte proliferation, cytolytic activity, and binding to target cells (Nemoto, Yu, and Dennert 1996;Wang, Nemoto, and Dennert 1996; Wang et al. 1994). In addition, a mADPRT has been implicated in the susceptibility of mice to two autoimmune diseases (Fowell and Mason 1993;Matthes et al. 1997). In Phycomyces blakesleanus, an endogenous mADPRT regulates glyceraldehyde-3-phosphate dehydrogenase activity in vitro; while ADP-ribosylation inhibitors diminished spore germination of this fungus (Deveze-Alvarez, Garcia-Soto, and Martinez-Cadena 1997;Deveze-Alvarez, Garcia-Soto, and Martinez-Cadena
2001). Bacteria utilize mADPRT as a mechanism of virulence and as a way to control important physiological processes, such as nitrogen fixation (Ludden 1994).In the first case are the potent bacterial toxins that interfere with cellular functions by catalyzing the mADP-ribosylation of key regulatory proteins in human hosts (Rappuoli and Pizza 2000). Some bacteria secrete soluble mADPRT toxins that have receptors on the surface of target cells and are able to modify specific intracellular acceptor proteins. Examples are the Diphtheria toxin and Pseudornonas exotoxin A, which ADP-ribosylate elongation factor 2; and Escherichia coli LT1 and cholera toxins, whose acceptor is a Gcx protein (Staddon, Bouzyk, and Rozengurt. 1991).Other bacterial ADP-ribosylating toxins are secreted by a type I11 export Corresponding Author: E. Avila-Telephone
number: 52 (473) 732
43 02, ext. 8174; FAX number: 52 (473) 732 43 02 ext. 8153. E-mail:
[email protected]
pathway and enter directly to the target cell (Yahr, Goranson, and Frank 1996). In pathogenic protozoan, such as Entamoeba histolytica, the existence of mADPRT has not been described as a virulence mechanism or as a way to regulate its own physiological processes. Entamoeba histolytica trophozoites are highly phagocytic and cytotoxic cells, causing severe public health problems in developing countries. Entamoeba’ s pathogenicity seems to involve a complex set of aggressive factors such as adhesins, amebapores, and proteases (Gilchrist and Petri 1999; Stanley and Reed 2001). In this work we describe an extracellular mADPRT activity with putative specificity for the amino acid arginine. Because of its extracellular localization, it may play an important role in parasite-host interaction. MATERIALS AND METHODS
Entamoeba histolytica cultures. Trophozoites of Entarnoeba histolytica HMI strain were grown in TYI-S-33 medium with 15% bovine serum and 1% Diamond vitamin mixture (Diamond, Harlow, and Cunnick 1978). Trophozoites were harvested by centrifugation at 72 h of culture, at the end of the logarithmic growth phase. Cells were suspended in TYI medium, without serum and centrifuged at 250 g for 4 min at 4-8 O C .
Entamoeba histolytica extracellular medium. Extracellular material of E. histolytica was obtained by incubating 5 X lo6 trophozoites/ml for 1 h at 37 “C in TYI medium, without serum, containing 50 FM trans-epoxysuccinyl-1-leucylamido(4-guanidino)-butane (E-64)as cysteine protease inhibitor. In some experiments, Brefeldin A (50 Fg/ml) was also added to TYI medium. After incubation, trophozoites were carefully removed by centrifugation at 250 g for 4 min. The viability of the trophozoites was at least 95%. Any cellular debris in the extracellular medium was removed by centrifugation at 3,000 g for 15 min. This ameba-free medium was concentrated by lyophylization. The effect of Brefeldin A on the secretion of ADP-ribosyl transferase activity (enzyme and substrate) was assayed as follows: fresh TYI-S-33 culture media, with or without 50 Fg/ml of Brefeldin A, were preheated at 37 “Cfor 15 min. These fresh media were then added under sterile conditions to culture tubes containing adherent trophozoites grown for about 64 h in which their own media were decanted. Cell culture was continued for 2-8 h. Finally, trophozoites from each condition were harvested
454
DELGADO-CORONA ET AL.-EXTRACELLULAR
and extracellular media were obtained as described above in the presence or absence of 50 pg/ml of Brefeldin A. As Brefeldin A could interfere with secretion of proteins other than the mADPRT, the protein amount was not equalized in each gel lane in this experiment, instead equal vol. of extracellular ameba-free medium were used. These were derived from trophozoites incubated exactly at the same cellular density and under identical conditions (except for the presence or absence of Brefeldin A). ADP-ribosyl transferase assay. The composition of the standardized ADP-ribosylation reaction mixture (100 pl) was as follows: 50 mM Tris-HCI, pH 7.0, 1 mM DTT, 2 p.M NAIF. 0.5 mM ATP, 2 mM MgCI,, 80 pM E-64, 2 pCi ["PINAD(1,000 Ci/mmol), or 2 pCi ['4CC-carbonyl]NAD' (51 mCi/ mmol), and 100 pg of protein equivalents from the extracellular medium as the source of both enzyme and substrate(s). As [:'CnicotinamidejNAD' was of a lower specific activity than the [32P]NAD+,non-radioactive NAD' was omitted from reactions utilizing ['4CC-carbonyl]NAD+. Reaction mixtures were incubated for 1 h at 30 "C. In some experiments, non-radioactive NAD+, or non-radioactive ADP-ribose were added and the assay performed under the same conditions as above. Reactions were stopped by the addition of 25 4 of sample buffer for electrophoresis 5 X, without 2-mercaptoethanol. Samples were boiled for 3 min to inactivate proteases; 2-mercaptoethanol was then added to a 5% final concentration and the samples were boiled again for 1 min before being analyzed by electrophoresis on 10% SDS-polyacrylamide gels (Laemmli 1970), stained with Coomassie blue, heat-dried, and then exposed for 3 to 7 d to X-Omat films (Kodak, Cat. No.1900984). Enrichment of ADP-ribosyl transferase activity. As initial steps for the future purification of the ameba mADPRT, the extracellular medium was centrifuged at 100,000 g for 1 h. The supernatant contained all the enzymatic activity and 90% of the total TCA precipitable protein. The extracellular ultracentrifuged medium was sequentially precipitated with 40, 60, and 80% saturated ammonium sulfate (4 "C). The 60% saturated ammonium sulfate precipitate contained the mADPRT activity. This was redissolved in PBS with 10 pM E-64 and fractionated further by size exclusion chromatography on Bio-Gel P200. The resulting fractions were assayed to detect the mADPRT activity and its substrates. Stability of the binding between ADP-ribose and the acceptor amino acid. The ADP-ribosylation reactions using enriched extracellular mADPRT as the source of enzyme and substrate were performed as indicated above, except five times more reaction mixture (500 pl) was prepared. After incubation, the reaction mixture was divided into five equal aliquots. The first aliquot (control) was treated with sample buffer for electrophoresis and maintained at 4 "C for electrophoretic analysis. The other four aliquots were treated at room temperature for 1 h with the following reactants: 1 M NH,OH, pH 7.0; 1 mM HgCl?, 750 mM Tris-HCI, pH 9.0; or 300 mM Tris-HCI, pH 7.0 respectively, each at the indicated final concentration. After these treatments, samples were analyzed by autoradiography of the gel as indicated above. All experiments were performed at least twice. Representative results are shown.
mADP-RIBOSYL TRANSFERASE IN AMEBA
455
Fig. 1. Effect of the absence of individual components of the reaction rnikrur? D I ~,\DP-rihosylation by the extracellular medium of Entc;uioehLi iii_cio!:.ricc. .XDP ribosylation was performed at 30 "C for 1 h, in 100 pl of rh? fcillo\\ inp reaction mixture (lane 1): SO mM Tiis-HC1, pH 7.0. 1 n i l 1 DTT. p\l S;\D-. 0.5 mM ATP, 2 mM MgCI?, 80 p M E-61.7 pCi [-rP]S.AD- and 100 pp of extracellular ameba-free medium, as source of rnr!nie and >ubstrates. In lanes 2 to 6 the following components \yere indil iduall! omitted in the ADP-ribosylation reaction: DTT (lane 2 ) . ATP (lane 3 ) . Tris-HC1 (lane 4), non-radioactive NAD' (lane 5), and MgCl, (lane 6). Lane 0--molecular weight markers. A) Coomassie-blue stained gel: B ) autoradiograph.
ribosylated band, corresponding to an Mr 37,000 polypeptide, was observed in the resulting autoradiographs (Fig. lB). Two minor radiolabeled products with molecular weights of 1 1 1 and 97 kDa were also observed (Fig. 1 B). This extracellular activity was not due to cell lysis because trophozoite's viability was 95% in all the experiments, an intracellular ameba protein (Rho) was not detected in the extracellular medium and the yield of labeled products was diminished using medium derived from trophozoites that had been incubated with Brefeldin A (see later). In addition, an amount of protein from a total extract of trophozoites equivalent to 5% of lysed cells did not produce radiolabeled products at the same experimental conditions. In total extracts from E. histolytica trophozoites, several radiolabeled products were observed in autoradiographs exposed for 7 d (not shown), in contrast to the major radiolabeled 37 kDa band of the unfractionated extracellular medium. Culture medium (TYI without serum) is not the source of enzyme or substrates since, when TYI alone was used in the reaction, no bands were observed either in the Coomassie-stained gel or in the autoradiographs (data not shown). To optimize the reaction conditions, several components of the reaction mixture were omitted, one at each reaction. As observed in Fig. 1B (lane 2), the absence of DTT reduced the label in the substrate, as compared with Fig. 1B (lane l), where all the components were present. The effect of DTT was dosedependent: a maximum increase in the intensity of the 37-kDa band was observed at 5 mM of this compound (data not shown). The absence of ATP also reduced the label in the 37-kDa substrate (Fig. lB, lane 3). Tris-HCI, pH 7.0, seems to be an appropriate buffer for the reaction (Fig. lB, lane 4) while the absence of MgCI, had very little effect upon the intensity of the band (Fig. l B , lane 6). When the non-radiolabeled NAD+ was omitted the intensity of the 37-kDa band increased (Fig. lB, lane 5 ) , indicating a dilution of [3'P]NAD+ in the control reaction (Fig. lB, lane I), where 2 pM non-radioactive NAD+ RESULTS was present. As a control, the Coomassie-stained gel is shown: Detection of an extracellular mADP-ribosyl transferase similar protein patterns are observed in each lane (Fig. lA, activity in E. histolytica. On the bases that the putative enzyme lanes 1 to 6). In order to establish the best non-radioactive NAD' concenmay utilize a substrate in host target cells or act on its own substrate(s), ameba-free extracellular TYI medium derived from tration for the reaction, different concentrations were tested a 1-hour 37 "C incubation of E. histolytica trophozoites was from 0 to 100 pM. As expected, if the non-radioactive NAD+ used as a source of enzyme and substrate(s). A major [12P]ADP- concentration increased, further dilution of [32P]NAD+ OC-
45 6
J. EUKARYOT. MICROBIOL., VOL. 49, NO. 6, NOVEMBER-DECEMBER 2002
Fig. 2. Absence of Rho protein in the extracellular medium from Entamoeba histolyica. Total homogenate (lanes 2, 4) and extracellular medium (lanes 1,3) from E. hisrolytica were incubated in ADP-ribosylation conditions with (lanes 1, 2) and without (lanes 3, 4) 15 ng of C3 exoenzyme from Clostridium botulinum. After reaction, samples were analyzed in 12.5% polyacrylamide gels at 50 pg of protein per lane, except in lane 4 where 100 K g were loaded. Lane 0-molecular weight markers A) Coomassie-stained gel; B) autoradiograph.
curred, diminishing the intensity of radiolabeled product. At 1, 5 and 10 p,M NAD+, the major 37-kDa band was still clearly observed, but this radiolabeled band was almost lost at 100 pM NAD’ (data not shown). The [32P]ADP-ribosylationwas dependent upon the time of the reaction. The 37-kDa radiolabeled substrate was clearly detected after only 20 min of incubation, and the intensity of this band increased at longer incubations times (data not shown). Also, the optimum pH and temperature were determined: a [32P]ADP-ribosylatedband was observed at 30 and 37 “C, while the intensity of this band decreased at 25 or 40 “C. The enzyme was inactivated by incubation for 30 min at 55 “C. The optimum pH of the reaction was 7.0; at pH 7.4 or higher the intensity of the radiolabeled band diminished drastically (data not shown). Rho, an intracellular protein, is undetectable in ameba extracellular medium. C3 exoenzyme from Clostridium botulinum specifically ADP-ribosylate Rho proteins, including Rho from E. histolytica (Godbold and M a n 2000). To exclude the contribution of cell lysis (5%) on the extracellular localization of the ADP-ribosyl transferase, it was determined if an intracellular protein was present in ameba extracellular medium. The ameba protein Rho was identified by ADP-ribosylation with the C3 exoenzyme. As shown in Fig. 2, two ADP-ribosylated bands of 25 and 22 kDa were observed in total homogenate from E. histolytica in the presence of C3 exoenzyme (Fig. 2B, lane 2) but not in its absence (Fig. 2B, lane 4). In the extracellular medium, C3 exoenzyme did not ADP-ribosylate any band (Fig 2B, lane l), only the ameba 37 kDa band was observed in presence or absence of C3 in this autoradiograph exposed for 3 d (Fig. 2B, lanes 1, 3). At longer exposition times (7 d) several endogenous ADP-ribosylated bands were observed in the total homogenate from ameba. In Fig. 2A, lane 4, 100 bg of protein from total homogenate were loaded instead of 50 bg as in the other lanes (Fig. 2A, lanes 1-3) to ensuring that no endogenous ADP-ribosylated bands could be confused with the Rho protein ADP-ribosylated by C3 exoenzyme. Brefeldin A effect on the secretion of mADP-ribosyl transferase. The fungal toxin Brefeldin A dissociates coat proteins from Golgi membranes and causes the rapid disassembly
Fig. 3. Effect of Entamoeba histulytica preincubation with Brefeldin A on the “secretion” of the ADP-ribosybdtion system, enzyme and substrate. Equal volumes of extracellular media were employed for ADP-ribosylation. The media were obtained from the same ameba densities under identical conditions except for the presence or absence of Brefeldin A. Lane 1 (A, B)-ADP-ribosylation of control extracellular medium. Lane 2 (A, B)-ADP-ribosylation in extracellular medium obtained from trophozoites preincubated with 50 Kg/ml Brefeldin A. Lane 3 (A, Bhreversal of Brefeldin A effect, following incubation without Brefeldin A for 2 h. Lane &molecular weight markers. A) Coomassiestained gels; B) autoradiograph.
of the Golgi complex, inhibiting secretion via vesicular transport (Klausner, Donaldson, and Lippincott-Schwartz 1992; Lippincott-Schwartz et al. 1989). A similar effect has been observed on Entamoeba histolytica (Ghosh et al. 1999). Ameba trophozoites were pre-incubated under normal culture conditions with 50 kg/ml Brefeldin A for 8 h. They were then harvested and used to obtain extracellular material during 1-h incubation in the presence of Brefeldin A. The yield of [32P]ADPribosylated 37-kDa product was markedly diminished using medium derived from cells preincubated with the inhibitor (Fig. 3B, lane 2). Brefeldin A also affected the secretion of E. histolytica proteins as observed in Fig. 3A (lane 2). Brefeldin A did not inhibit the ADP-ribosylation reaction itself since no effect was observed when it was added directly to the assay (data not shown). To determine whether the Brefeldin A effect was reversible, as reported in other cells (Klausner, Donaldson, and Lippincott-Schwartz 1992), the inhibitor was removed from the culture medium after 7 h of treatment. Trophozoites culture was continued without Brefeldin A for 1 h. Extracellular medium was then obtained as usual for 1 h in TYI without Brefeldin A. The ADP-ribosylation activity was regained when Brefeldin A was removed (Fig. 3B, lane 3), as compared with sample where Brefeldin A remained in the medium (Fig. 3B, lane 2). Partial purification of the ADP-ribosyl transferase and its substrates. In order to purify the secreted mADPRT, an initial enrichment of the activity was attempted. First, the extracellular medium was fractionated by treatment with ammonium sulfate at 40, 60 and finally 80% saturation. Most of the activity was recovered in the precipitate formed at 60% saturation. When this fraction was used in the enzyme reaction, several radiolabeled products were observed (Fig. 4B, lane 3). The 37-kDa substrate that initially was the major radiolabeled substrate was also present but with less intensity than the new substrates. The 60% saturated ammonium sulfate precipitate was fractionated further by size exclusion chromatography. The enzyme activity and the substrates eluted together, mainly in fractions 14 and 15 (Fig. 4D), these fractions correspond to molecular weights of 105 to 91 kDa. When mixtures of different fractions were
DELGADO-CORONA ET AL.-EXTRACELLULAR
mADP-RIBOSYL ‘RANSFERASE IN AMEBA
457
Fig. 5. Lack of competition of ADP-ribose with the ADP-ribosylation system of Enturnoeba histolyrica. ADP-ribosylation reactions as in Fig 1 (control),but using the partially purified extracellular mADPRT (100 pg protein per lane), and 0 (lane l), 50 (lane 2), 100 (lane 3), and 150 p M (lane 4) of non-radioactive ADP-ribose. Lane 0-molecular weight markers. A) Coomassie-stained gel; B) autoradiograph.
Fig. 4. Partial purification of ADP-ribosylation system, enzyme and substrates from extracellular medium of Enrnmoebn histolyticu. In A and B, extracellular medium (lane 1) was sequentially precipitated with 40% (lane 2), 60% (lane 3), and 80% (lane 4) saturated ammonium sulfate. Samples were dialyzed against Tris-HC1 10 mM, pH 7.0, containing E-64, 8 pM, before ADP-ribosylation reactions. No dialyzed 60% precipitate was separated by size exclusion chromatography, fractions 13 to 16 in panels C) and D) showed the highest ADP-ribosylation activity. One-hundred micrograms protein were loaded per lane. A) and C ) Coomassie-stained gels; B) and D) autoradiographs.
made, new radiolabeled substrates were not observed (data not shown). Specificity of the mADP-ribosylation reaction. To determine if ameba samples contained NADglycohydrolase activity that might have split [32P]NAD+ and generated [32P]ADP-ribose, which then could bind non-enzymatically to ameba substrates, competition experiments were performed. MonoADPribosylation reactions were done in the absence or in the presence of increasing concentrations of non-radioactive ADP-ribose, using crude extracellular medium and the partially purified extracellular mADPRT as source of enzyme and substrates. No competition was observed with any of the ADPribose concentrations used, neither in the crude extracellular medium (data not shown) nor in the partially purified extracellular mADPRT (Fig. 5B, final concentrations of ADP-ribose 0, 50, 100 and 150 pM lanes 1 4 , respectively). It has been reported (Itoga et al. 1997) that in ADP-ribosylation reactions, the glyceraldehyde-3-phosphatedehydrogenase (GAPDH) from yeast is labeled as a result of non-enzymatic binding of the complete NAD molecule to cysteine residues. To exclude this possibility in ameba samples, monoADP-ribosylation assays were made with [“T-carbonyl]NAD+ and [32P]NAD+,using the partially purified extracellular mADPRT and the yeast GAPDH as a control. GAPDH was labeled as
expected with both [14C-carbonyl]NAD+(Fig. 6B, lane 4) and with [32P]NAD+(Fig. 6B, lane 2). In ameba samples, substrate labeling occurred only with [32P]NAD’ (Fig. 6B, lane 1) but not with [I4C-carbonyl]NAD+ (Fig. 6B, lane 3), excluding the possibility of non-enzymatic binding of complete NAD’ molecules to ameba substrates. Stability of ADP-ribose binding to ameba substrates. To determine the nature of the ADP-ribose linkage to ameba substrates, [32P]ADP-ribosylatedproteins were exposed to neutral 1 M NH,OH, 1 mM HgCl,, or pH 9.0, conditions known to cleave the linkage of ADP-ribose to arginine, thiol, and carboxylate groups, respectively. The 37-kDa substrate in the extracellular medium (data not shown) and the substrates observed in the partially purified mADPRT fraction were resistant to HgC1, (Fig. 7B, lane 3) and pH 9.0 (Fig. 7B, lane 4), but they were sensitive to 1 M NH,OH, pH 7.0 (Fig. 7B, lane 2). DISCUSSION MonoADP-ribosyl transferases have been implicated in key physiological and pathological events, but they have not been
Fig. 6 . “ADP-ribosylation” with [‘?PINAD+ and [I4C-carbonyl]NAD+ with the enriched extracellular mADPRT from Enttimoeba histolyticu and with glyceraldehyde-3-phosphate dehydrogenase. “ADP-ribosylation reactions” were performed with yeast GAPDH (5 pg per lane, lanes 2 and 4) and the partially purified extracellular mADPRT at 100 pg protein per lane (lanes 1 and 3), as source of enzyme and substrates. [32P]NAD+was used in lanes 1 and 2 and V4Ccarbonyl]NAD+ in lanes 3 and 4. Lane 0-molecular weight markers. A) Coomassie-blue stained gel; B) autoradiograph.
458
J. EUKARYOT. MICROBIOL., VOL. 49, NO. 6, NOVEMBER-DECEMBER 2002
Fig. 7. Stability of the binding between ADP-ribose and the acceptor amino acid in the ADP-ribosylation system of Entunzoebu histulyfica. ADP-ribosylation was performed in 50O-pl reaction mixture with the partially purified extracellular mADPRT. After the reaction had proceeded for 60 min, the mixture was divided into five tubes of 100 p1 each (100 pg protein per lane). Tube I (lane 1) received no additional treatment, tubes 2 to 5 were treated 1 h at room temperature with the following: 1 M NH,OH, pH 7.0, (lane 2); I mM HgC12, (lane 3); 750 mM Tris-HC1, pH 9.0 (lane 4); and 300 mM Tris-HC1, pH 7.0 (lane 5). Lane 0-molecular weight markers. Panel A) Coomassie blue-stained gel; B) autoradiograph.
reported in protozoan parasites. In this study, a mADPRT activity and three substrates were observed in unfractionated extracellular medium obtained by incubating E. histolytica trophozoites for 1 h at 37 "C. The major substrate has an apparent molecular weight of 37 kDa. Its labeling was dependent on time, temperature, pH, the presence of ATP, DTT, and the ratio between non-radioactive NAD' and [32P]NAD+in the assay, supporting its dependence upon an enzymatic reaction. After two initial purification steps, several new substrates were observed in the ameba extracellular medium, perhaps due to their enrichment. Another possibility is that as the initial extracellular TYI medium contained several peptide components, they may compete as acceptors for ADP-ribosylation. It is also possible that new arginine residues are exposed in the proteins as a result of ammonium sulfate precipitation, allowing their ADP-ribosylation. In fact, some monoADP-ribosyl transferases could modify other proteins different to their physiologic acceptors when used in cell extracts. It has been reported (Gill and Meren 1978; Staddon, Bouzyk, and Rozengurt 1991) and also we have observed that cholera toxin, an arginine-specific mADPRT, ADP-ribosylate many proteins in vitro, in spite of it has only one physiologic acceptor. Other mADPRT such as the C3 toxin from Clostridium botulinum are very specific and only modify its physiologic acceptor even if they are used in complex samples of proteins. In this study, the C3 exoenzyme from Clostridium botulinum was used as a tool to identify the ameba Rho protein in the total homogenate from ameba. In fact, two bands were ADPribosylated (25 and 22 kDa) by C3 exoenzyme, in accord with a previous report (Godbold and Mann 2000). Rho protein was not detected in the extracellular medium supporting the view that the rnADPRT present in the extracellular medium is not the result of ameba lysis. Other evidences that the extracellular mADPRT did not originate from the lysed trophozoites are: the absence of mADP-ribosylated bands in a total extract of trophozoites equivalent to the 5% of lysed cells and the distinct patterns of ADP-ribosylated bands in total extract (data not shown) compared with the unfractionated extracellular medium. The effect of Brefeldin A on the detection of extracellular
ADP-nbosylated 37-kDa substrate also suggests that ameba mADPRT could be a real secreted enzyme or a membrane protein that depends upon secretion for insertion on the ameba cell surface. In the latter case, the enzyme should somehow be released from the plasma membrane. In fact, several GPI-linked mADPRT, which are released from the cell surface, have been described in eukaryotic cells (Okazaki and Moss 1998). The mouse ART2.2 is released by a metalloprotease from the surface when T cells are activated, (Kahl et al. 2000). Presumably, GPI-anchored mADPRT could also be released by a GPI-specific phospholipase C. The kinetics of "secretion" of mADPRT was not determined during Entamoeba histolytica growth due to interference with the 15% bovine serum present in the ameba culture. However, the enzyme was released at 37 "C in TYI medium without serum, in the first 5 min and increasing until 60 min of ameba incubation (data not shown). At this time ameba trophozoites were 95% viable in the medium without serum. This kinetics was similar to the releasing of mouse ART2.2-50% of this enzyme is released from cell surface within 15 min (Kahl et al. 2000). As yeast GAPDH reacts non-enzymatically to the complete NAD molecule (Itoga et al. 1997), it was important to exclude this possibility in the observed ameba substrates. The lack of substrate labeling with [14C-carbonyl]NAD+and the absence of competition with non-radioactive ADP-ribose exclude non-enzymatic labeling of substrates, indicating that there is at least one mADPRT activity from E. histolytica in extracellular medium. On the basis of sensitivity to NH,OH it is strongly suggested that mono-ADP-ribosylation occurred at arginine residues on the ameba substrates. If more than one activity was present in the enriched mADRT, they may have the same specificities for arginine residues in the substrate proteins. This evidence has not been corroborated by using agmatine or poli-L-arginine, because amebic extracellular monoADP-ribosyl transferase had not shown activity on these substrates. It is possible that the enzyme might recognize an acceptor protein with a motif of several amino acids to modify arginine, or more likely that the enzyme in its present unpurified state is competed for by so many protein substrates that it is unable to modify the more simple ones, such as agmatine or poli-L-arginine. The unfractionated extracellular medium from ameba did not show clear ADP-ribosylation of histones from calf thymus, but the precipitate obtained with 60% saturation of ammonium sulfate from the extracellular medium showed ADP-rybosylation of a histone band (data not shown) supporting the likely ADP-ribosylation of an arginine residue. The evidence presented in this paper clearly indicates that E. histolytica has both an extracellular mADP-ribosyl transferase and its substrates. This enzyme of E. histolytica is of interest for its possible role in regulatory functions and/or its role in virulence mechanisms, if the enzyme has any substrates in a target cells. ACKNOWLEDGMENTS This study was supported by research CONACYT grants no. 28194-N, and 35269-N. PDC and AHA were supported by postgraduate scholarships from CONACYT, Mexico.
LITERATURE CITED Deveze-Alvarez, M., Garcia-Soto, J. & Martinez-Cadena, G. 1997. Molecular characterisation of a fungal mono(ADP-ribosy1)trmsferase. Adv. Exp. Med. B i o l , 419:155-162 Deveze-Alvarez, M., Garcia-Soto, 3. & Martinez-Cadena, G. 2001. Glyceraldehyde-3-phosphatedehydrogenase is negatively regulated
DELGADO-CORONA ET AL.-EXTRACELLULAR by ADP-ribosylation in the fungus Phyconryces blcikesleeanus. Microbiology, 1472579-2584.
Diamond, L. S., Harlow, D. R. & Cunnick, C. C. 1978. A new medium for the axenic cultivation of Entamoeba hisrolxtica and other Emamoeba. Trans. R. Soc. Trop. Med. Hyg., 72:431132. Ehret-Hilberer, S., Nullans, G., Aunis, D. & Virmnu.;. S. 1992. \lono ADP-ribosylation of transducin catalyzed b> rod outer .egment extract. FEES Lett., 309:394-398. Fowell, D. & Mason, D. 1993. Evidence that the T csll repertoke of normal rats contains cells with the potenrial to cause diabetes. Chxacterization of the CD4+ T cell subset that inhibits thi.; zutoimmun? potential. J. Exp. Med., 177:627-636. Ghosh, S. K., Field, J., Frisardi. M.. Rosenthal. B.. \ h i . Z . . Roger,. R. & Samuelson, J. 1999. Chitinase secretion b!. enc? sting Ditcimoeba invadens and transfected Eiztanzoeba hirtolxrica trophozoites: localization of secretory vesicles, endoplasmic reticulum. and Golgi apparatus. lnjkct. Zmmun., 67:3073-308 1. Gilchrist, C . A. & Petri, W. A. 1999. Virulence factors of Entamoeba histolytica. Curr. Opin. Micr-vbiol., 2:433-437. Gill, D. M. & Meren, R. 1978. ADP-ribosylation of membrane proteins catalyzed by cholera toxin: basis of the activation of adenylate cyclase. Proc. Natl. Acad. Sci. USA, 75:3050-3054. Godbold, G. D. & Mann, B. J. 2000. Cell killing by the human parasite Enramoeba histolyfica is inhibited by the Rho-inactivating C3 exoenzyme. Mol. Biochem. Purasitol., 108:141-15 1. Itoga, M., Tsuchiya, M., Ishino, H. & Shimoyama, M. 1997. Nitric oxide-induced modification of glyceraldehyde-3-phosphatedehydrogenase with NADf is not ADP-ribosylation. J. Biochem. (Tokyo), 121:1041-1046. Kahl, S., Nissen, M., Girisch, R., Duffy, T., Leiter, E. H., Haag, F. & Koch-Nolte, F. 2000. Metalloprotease-mediated shedding of enzymatically active mouse ecto-ADP-ribosyltransferase ART2. 2 upon T cell activation. J. Iiiznzunol., 165:4463-4469. Kharadia, S. V., Huiatt, T. W., Huang, H. Y., Peterson, J. E. & Graves, D. J. 1992. Effect of an arginine-specific ADP-ribosyltransferase inhibitor on differentiation of embryonic chick skeletal muscle cells in culture. Exp. Cell. Rex. 201:33-42. Klausner, R. D., Donaldson, J. G. & Lippincott-Schwartz, J. 1992. Brefeldin A: insights into the control of membrane traffic and organelle structure. J. Cell. Biol., 116:1071-1080. Laemmli, U. K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature, 227:680-685. Lippincott-Schwartz, J., Yuan, L. C., Bonifacino, J. S. & Klausner, R. D. 1989. Rapid redistribution of Golgi proteins into the ER in cells treated with brefeldin A: evidence for membrane cycling from Golgi to ER. Cell, 56:801-813. Ludden, P. W. 1994. Reversible ADP-ribosylation as a mechanism of enzyme regulation in procaryotes. Mol. Cell. Biochem., 138:123-129. Lupi, R., Corda, D. & Di Girolamo, M. 2000. Endogenous ADP-ribo-
mADP-RIBOSYL TRANSFERASE IN AMEBA
459
sglation of the G protein beta subunit prevents the inhibition of type 1 adenylyl cyclase. J. Biol. Chem., 275:9418-9424. Matthes. M.. Hollmann, C . , Bertuleit, H., Kuhl, M., Thiele, H. G., Haag, F. CQ Koch-Nolte, F. 1997. “Natural” RT6-1 and RT6-2 “knockout” mice. .4di.. ESP. Med. Biol., 419:271-274. \loss. J.. Jacobson, M. K. & Stanley, S. J. 1985. Reversibility of arginine-specific mono(ADP-ribosy1)ation: identification in erythrocytes of an .lDP-iibose-L-arginine cleavage enzyme. Proc. Natl. Acnd. Sci. i-S.4. 825603-5607. \fuss. J.. Stank?. S . J.: Nightingale, M. S., Murtagh, J. J., Monaco, L., Iriishirna. I(..Chcn. H. C., Williamson, K. C . & Tsai, S. C . 1992. Sfolecular and immunological characterization of ADP-ribosylarginine h!-drolasss. J. Biol. Clzem., 267:10481-10488. \Ioss. J.. Balducci. E.. Cavanaugh, E., Kim, H. J., Konczalik, P., Lesma, E. A.. Okazaki. I,. J.. Park. M., Shoemaker, M., Stevens, L. A. & Zolkiewska. X. 1999. Characterization of NAD:arginine ADP-ribosylrransferases. Mol. Cell. Biochem., 193:109-1 13. Nemoto. E.. 1-u.Y.gi Dennert. G. 1996. Cell surface ADP-ribosyltransferase regulates lymphocyte function-associated molecule-1 (LFA-1) function in T cells. J. I m m i ~ d . 157:3341-3349. , Okazaki, I. J. & Moss. J. 1998. Glycosylphosphatidylinositol-anchored and secretory isoforms of mono-ADP-ribosyltransferases. J. Biol. Chem., 27323617-23620. Rappuoli, R. & Pizza, M. 2000. Bacterial toxins. In: Cossart, P., Boquet, P., Normark, S. & Rappuoli, R. (ed.), Cellular Microbiology. Washington, D. C. p. 193-220. Saari, L. L., Pope, M. R., Murrell, S . A. & Ludden, P. W. 1986. Studies on the activating enzyme for iron protein of nitrogenase from Rhodospirillum rubrum. J. Biol. Chem., 261:4973-4977. Staddon, J. M., Bouzyk, M. M. & Rozengurt, E. 1991. A novel approach to detect toxin-catalyzed ADP-ribosylation in intact cells: its use to study the action of Pastcurella multocidu toxin. J. Cell. Biol., 115:949-958. Stanley, S. L., Jr.& Reed, S. L. 2001. VI. Entamoeba histolytica: parasite-host interactions. Am. J. Physiol. Gastrointest. Liver Physiol., 280:G 1049-1 054. Wang, J., Nemoto, E. & Dennert, G. 1996. Regulation of CTL by ectonicotinamide adenine dinucleotide (NAD) involves ADP-ribosylation of a p56lck-associated protein. J. Immunol., 156:28 19-2827. Wang, J., Nemoto, E., Kots, A. Y., Kaslow, H. R. & Dennert, G. 1994. Regulation of cytotoxic T cells by ecto-nicotinamide adenine dinucleotide (NAD)correlates with cell surface GPI-anchoredlarginine ADP-ribosyltransferase. J. Imrnunol., 153:4048-4058. Yahr, T. L., Goranson, J. & Frank, D. W. 1996. Exoenzyme S of Pseudomonas aeruginosa is secreted by a type 111 pathway. Mol. Microbiol., 22:991-1003. Yuan, J., Huiatt, T. W., Liao, C . X. Robson, R. M. & Graves, D. J. 1999. The effects of mono-ADP-ribosylation on desmin assemblydisassembly. Arch. Biuchenz. Biophys., 363:314-322. Received 03/08/02. 09/01/02; accepted 09/02/02
