Proteolytic activation of internalized cholera toxin within hepatic endosomes by cathepsin D Cle ´ mence Merlen*, Domitille Fayol-Messaoudi*, Sylvie Fabrega, Tatiana El Hage, Alain Servin and Franc¸ois Authier Institut National de la Sante ´ et de la Recherche Me ´ dicale U510, Faculte ´ de Pharmacie Paris XI, Cha ˆ tenay-Malabry, France Cholera toxin (CT) produced by Vibrio cholerae is the major virulence factor responsible for the massive secretory diarrhea of infected humans [1]. CT interacts with intestinal epithelial cells and induces chloride secretion due to toxin-mediated activation of adenylate cyclase and elevation of intracellular cAMP. Activation of adenylate cyclase results from ADP-ribosylation of Arg201 of the a-subunit of the stimulatory GTP-binding regulatory protein, Gsa, catalyzed by the toxin [2]. CT ( 84 kDa) is an oligomeric protein of the A-B type composed of one activating A subunit (CT-A, r.m.m. 27 400 Da) and five identical B subunits (CT-B, r.m.m. 11 600 Da) arranged in a ring-like configur- ation that bind ganglioside GM1 at the cell surface [3]. The CT-A subunit is comprised of two functional domains termed the A1 and A2 peptides linked by a single disulfide bond. The A1 peptide exhibits the toxin’s ADP-ribosyltransferase activity which is neces- sary for CT cytotoxic action. The A2 peptide contains the endoplasmic reticulum-targeting motif KDEL at its C-terminus. For full ADP-ribosylation of the stimula- tory heterotrimeric GTPase Gsa, enzymatic production of a degradative fragment generated from native CT and structurally related to the A1 peptide must occur followed by its targeting to the Gsa substrate. This process, which takes 30–40 min in most cell types, Keywords cathepsin D; cholera toxin; endocytosis; hepatocyte; proteolysis Correspondence F. Authier, INSERM U510, Faculte ´ de Pharmacie Paris XI, 5 rue Jean-Baptiste Cle ´ ment, 92296 Cha ˆ tenay-Malabry, France Fax: +33 1 46835844 Tel: +33 1 46835843 E-mail: francois.authier@cep.u-psud.fr *These authors contributed equally to the paper (Received 29 April 2005, revised 1 July 2005, accepted 7 July 2005) doi:10.1111/j.1742-4658.2005.04851.x We have defined the in vivo and in vitro metabolic fate of internalized chol- era toxin (CT) in the endosomal apparatus of rat liver. In vivo, CT was internalized and accumulated in endosomes where it underwent degrada- tion in a pH-dependent manner. In vitro proteolysis of CT using an endo- somal lysate required an acidic pH and was sensitive to pepstatin A, an inhibitor of aspartic acid proteases. By nondenaturating immunoprecipita- tion, the acidic CT-degrading activity was attributed to the luminal form of endosomal cathepsin D. The rate of toxin hydrolysis using an endosomal lysate or pure cathepsin D was found to be high for native CT and free CT-B subunit, and low for free CT-A subunit. On the basis of IC 50 values, competition studies revealed that CT-A and CT-B subunits share a com- mon binding site on the cathepsin D enzyme, with native CT and free CT-B subunit displaying the highest affinity for the protease. By immuno- fluorescence, partial colocalization of internalized CT with cathepsin D was confirmed at early times of endocytosis in both hepatoma HepG2 and intestinal Caco-2 cells. Hydrolysates of CT generated at low pH by bovine cathepsin D displayed ADP-ribosyltransferase activity towards exogenous Gsa protein suggesting that CT cytotoxicity, at least in part, may be rela- ted to proteolytic events within endocytic vesicles. Together, these data identify the endocytic apparatus as a critical subcellular site for the accu- mulation and proteolytic degradation of endocytosed CT, and define endo- somal cathepsin D an enzyme potentially responsible for CT cytotoxic activation. Abbreviations CT, cholera toxin; EN, endosomes; HI, human insulin; PA, pepstatin-A; PDI, protein disulfide isomerase. FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS 4385 corresponds to the lag-phase required for CT to enter the cell by endocytosis using clathrin-dependent and -independent mechanisms [4,5], become activated and move to its site of action. Two major candidate compartments have been proposed as being physiologically relevant to the mechanism of activation of internalized CT. The first activating pathway, identified in rat hepatocytes, has been proposed to operate at an early stage of endo- cytosis within endocytic vesicles [6,7]. Thus, using a subcellular fractionation approach to address CT compartmentalization and activation in vivo, it was shown that activation of adenylate cyclase by CT in intact liver requires the association and subsequent processing of the toxin in an endosomal compart- ment. However, the nature of the relevant enzymatic activity (protease and ⁄ or reductase) as well as the mechanisms underlying CT-A subunit cytotoxic action towards its substrate were not investigated. The other well-characterized subcellular compart- ment involved in CT activation is the endoplasmic reti- culum (ER) where CT is transported in a retrograde manner via the Golgi apparatus by a KDEL-dependent mechanism [8]. After entry of CT into the ER, the ER-chaperone protein disulfide isomerase (PDI) is involved in facilitating the reduction of the internal disulfide bond of CT [9] and preparation of the A1 peptide for retro-translocation to the cytosol by the Sec61p channel [10]. In the present study, we have characterized the endosomal processing of endocytosed CT in rat liver and examined the sensitivity of toxin proteolysis to pH, protease inhibitors and immunodepletion pro- cedures using antibodies to well-defined endosomal proteases. In addition, using a cell-free ADP-ribosyla- tion assay, we have explored the potential physiologi- cal significance of the endosomal proteolysis of CT in its cytotoxic action. This has allowed us to characterize the role of the aspartic acid protease cathepsin D in the endosomal degradation of internalized CT. We show that hydrolysates of CT generated in vitro by pure cathepsin D displayed ADP-ribosyltransferase activity towards exogenous Gsa protein. Results In vivo endocytosis of native CT and CT-B subunit in rat liver The kinetics of in vivo uptake of native CT and CT-B subunit into endosomal fractions were first assessed (Fig. 1). Rats were administered an intravenous injec- tion of native CT or CT-B subunit (50 lg per 100 g body weight) and killed 5–90 min postinjection. Following preparation of hepatic endosomes the amount of internalized CT-A and CT-B subunits was determined by western blot analyses (Fig. 1A). A simi- lar time-dependent increase in CT-A and CT-B A B Fig. 1. Kinetics of appearance of CT-A and CT-B subunits in hepatic endosomes after ligand administration. (A) Endosomal frac- tions were isolated at the indicated times after the in vivo administration of native CT or CT-B subunit, and evaluated for their con- tent of both subunits by western blot analy- sis using the polyclonal anti-CT Ig. Each lane contained  50 lg of endosomal protein. Arrows to the right indicate the mobility of CT-A ( 28 kDa) and CT-B subunit ( 12 kDa). Molecular mass markers are indicated to the left of each blot. (B) Assess- ment of polyclonal antibody specificity towards CT and individual A- and B-subunits was performed using reducing and non- reducing SDS ⁄ PAGE followed by western blot analysis. Each lane contained 1 lgof peptide. Arrows to the right indicate the mobility of CT-A ( 28 kDa), CT-B ( 12 kDa), and A1 peptide ( 21 kDa). A2 peptide was not detected under these experimental conditions. Molecular mass markers are indicated to the left of each blot. Endosomal proteolysis of cholera toxin by cathepsin D C. Merlen et al. 4386 FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS subunits was observed in endosomal fractions 5–30 min after native CT injection (Fig. 1A, upper blot), after which the level of both subunits decreased. A slower rate of B-subunit endocytosis was observed in response to CT-B subunit administration and remained elevated up to 90 min postinjection (Fig. 1A, lower blot). Characterization of the polyclonal anti-CT antibody C3062 by western blot analysis of pure CT or individ- ual A- and B-subunits revealed a specificity for both subunits under nonreducing conditions (Fig. 1B, upper blot). Following chemical reduction of the A-subunit interchain disulphide bond, the rabbit anti-CT anti- body recognized the A1 peptide and B-subunit with no immunoreactivity towards the A2 peptide (Fig. 1B, lower blot). Proteolysis of CT within hepatic endosomes at acidic pH We next examined the ability of hepatic endosomes to degrade CT (Fig. 2). The luminal and membrane- bound distribution of endosomal CT-degrading activity was assessed by western blot analysis of CT digestions performed at various pH (Fig. 2A). EN fractions degraded native CT at pH 4 and 5, with degradation decreasing markedly at pH 6–7. The CT-B subunit was more efficiently degraded than the CT-A subunit (Fig. 2A, EN, lane pH 4). A 25-kDa proteolytic frag- ment of CT-A subunit was specifically generated at pH 4–5. Another CT-A product of 22-kDa was observed in hydrolysates at pH 6–7 and also in the untreated toxin (lane -). Subfractionation of hepatic endosomes (EN) into a soluble endosomal lysate (ENs) revealed a similar pattern of proteolysis to that observed for the EN fraction (Fig. 2A, ENs), suggest- ing that the majority of CT-degrading activity in endo- somes is soluble. A soluble endosomal extract (ENs) was then assessed by HPLC analysis for its ability to proteolyze native CT in vitro at acidic (pH 4 and 5) and neutral pH (pH 7) (Fig. 2B). Degradation of both A- and B-subunits was pH-dependent with maximal degradation obtained at A B Fig. 2. Effect of pH on the endosomal pro- cessing of native CT. (A) Total (EN  10 lg) and soluble endosomal fractions (ENs  1 lg) were incubated with 7 lg native CT at 37 °C for 60 min in 30 m M citrate-phos- phate buffer at the indicated pH. The incu- bation mixtures were then analyzed by western blotting using the polyclonal anti-CT antibody. The mobility of each CT subunit is indicated on the right (CT-A  28 kDa; CT-B,  12 kDa). (B) Representative RP-HPLC pro- files obtained by incubating native CT (1 l M) with ENs ( 1 lg) at 37 °C for 60 min in 175 m M citrate-phosphate buffer at the indicated pH. Shown are absorbance profiles at 214 nm. Untreated CT subunits had an elution time of 60 min (CT-B) and 63 min (CT-A) (HPLC profile CT). The endosomal proteins alone did not give any detectable peaks (results not shown). C. Merlen et al. Endosomal proteolysis of cholera toxin by cathepsin D FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS 4387 pH 4. Intermediate peptide peaks were only observed at pH 4 in addition to the undegraded subunits which had decreased in peak height (Fig. 2B). Catalytic properties of endosomal CT-degrading activity We evaluated the ability of individual CT-A and CT-B subunits to act as substrates for endosomal acidic pro- teases (Fig. 3). The rate of hydrolysis of individual subunits was determined using RP-HPLC analysis by following the disappearance of the parent peptide after a 2 h incubation with ENs at pH 4 (Fig. 3, lower pan- els). The rate of peptide hydrolysis was found to be very low with free CT-A subunit and high with free CT-B subunit. However, when native CT was used as a substrate the rates of proteolysis of both subunits pre- sent in the AB 5 pentamer were comparable to that of the individual CT-B subunit (Fig. 3, lower panel, left). The effect of various protease inhibitors on the aci- dic CT-degrading activity contained in hepatic endo- somes was next examined by western blot and HPLC analyses (Fig. 4). The proteolytic activity directed against both A- and B-subunits at pH 4 was inhibited by pepstatin-A (PA), an inhibitor of aspartic acid pro- teases (Fig. 4A, PA). No effect was observed with the cysteine-protease inhibitor E-64 or the metal-chelating agent EDTA. The 25-kDa proteolytic fragment of the CT-A subunit, whose level was maximal after a 3 h incubation, was not detected in the presence of pepsta- tin-A. Two minor diffuse CT-A fragments of 22- and 18-kDa were also observed after 3 h of incubation, however, their production was inhibited in the pres- ence of pepstatin-A (for the 18-kDa species) or EDTA (for the 22- and 18-kDa species). RP-HPLC analysis of CT digestions performed at pH 4 using the ENs fraction confirmed the inhibitory effect of pepstatin-A, with no significant effect observed using E-64 and the serine protease inhibitor PMSF (Fig. 4B,C). Identification of endosomal CT-degrading enzyme as cathepsin D The inhibition of CT-degrading activity by pepstatin- A, its low pH optimum and its presence in the endo- somal lumen as a soluble form suggested cathepsin D as a likely candidate for this activity. We therefore used well characterized polyclonal antibodies to mature cathepsin D and its proform [11,12] to deplete cathepsin D from ENs (Fig. 5A). Quantitative immuno- precipitation of cathepsin D using antibodies direc- ted against the mouse (R291) and human enzyme (M8147) removed greater than 88% of the endosomal proteolytic activity directed towards both subunits as assessed by RP-HPLC (Fig. 5B). Immunodeple- tion of ENs with antibodies to cathepsin B and its Fig. 3. Kinetics of proteolysis of native CT, CT-A subunit and CT-B subunit at pH 4 by a soluble endosomal lysate. Shown are representative RP- HPLC profiles resulting from the incubation of native CT, CT-A or CT-B (1 l M) with ENs ( 1 lg) at 37 °C for 120 min in 175 mM citrate-phos- phate buffer pH 4 (lower HPLC profiles). Shown are absorbance profiles at 214 nm. Elution profiles of native CT and individual CT-A and -B subunits are shown for comparison (upper HPLC profiles). Endosomal proteolysis of cholera toxin by cathepsin D C. Merlen et al. 4388 FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS proform [13] failed to remove the proteolytic activity (Fig. 5A). To strengthen the physiological relevance of our observations obtained in vitro with cell-free endosomes, we studied the subcellular localization of internalized CT or CTB-FITC and cathepsin D in hepatoma HepG2 and intestinal Caco-2 cells by immunofluorescence con- focal microscopy (Fig. 6). Cells were incubated with 1 lm native CT for 30 min (HepG2 cells) or CTB-FITC for 15 min (HepG2 and Caco-2 cells) when most of the internalized ligands would be located in the endosomes (Fig. 1A). CTB-FITC and monoclonal antibody D15-8 to the CT-B subunit (in green) demonstrated a highly punctate staining pattern reminiscent of vesicular com- partments. Costaining with antibody directed against mature and precursor cathepsin D enzyme (in red) revealed a partial colocalization (in yellow) of CT with the intracellular aspartic acid protease. Affinity-binding and degradation of native CT and individual CT subunits by cathepsin D Substrates of the same protease would be expected to compete with each other for the enzyme binding site. We therefore used a competition assay to evaluate the ability of native CT and the individual A- and B-subunits to inhibit degradation of the radiolabeled substrate 125 I-labelled Tyr A14 -HI by cathepsin D (Fig. 7A). As reported previously [12,14], the well- defined cathepsin D substrate HI was found to inhibit A C B Fig. 4. Effect of protease inhibitors on the proteolysis of native CT by hepatic endosomes. (A) EN fraction was incubated with 7 lg native CT at 37 °Cfor1or3hin30m M citrate-phosphate buffer pH 4 without (buffer) or with 1% Me 2 SO (DMSO), 3.5 lgÆmL )1 pepstatin-A (PA), 0.1 m M E-64 or 1 mM EDTA. At the end of the incubation, the samples were analyzed by western blotting using the polyclonal anti-CT Ig. Molecular mass markers are indicated to the left of each panel. Arrows indicate the mobility of the intact A ( 28 kDa) and B subunits ( 12 kDa). (B) Representative RP-HPLC profiles obtained following the incubation of native CT with ENs in the presence or absence of 3.5 lgÆmL )1 pepstatin-A. Shown are absorbance profiles at 214 nm. Elution profile of native CT is shown for comparison (HPLC profile CT). (C) ENs ( 1 lg) was incubated with 1 l M native CT at 37 °C for 60 min in 175 mM citrate-phosphate buffer pH 4 in the absence or pres- ence of 3.5 lgÆmL )1 pepstatin-A (PA), 1% Me 2 SO (DMSO), 1 lM E-64, 1 mM PMSF or 1% MeOH. At the end of the incubation, the proteo- lytic reaction was stopped with acetic acid (15%), and the incubation mixtures were analyzed by RP-HPLC. The rate of proteolysis of CT-A and CT-B subunits was determined by following the disappearance of the peak area corresponding to the parent peptides. C. Merlen et al. Endosomal proteolysis of cholera toxin by cathepsin D FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS 4389 125 I-labelled Tyr A14 -HI proteolysis by bovine cathepsin D in a dose-dependent manner with an IC 50 of 30 lm. Native CT was  50 times more effective than HI at competing for proteolysis of radiolabeled HI (IC 50 of 0.6 lm). The CT-B subunit (IC 50 of 0.9 lm) was able to inhibit the degradation of radiolabeled HI compar- ably to that of native CT, whereas the CT-A subunit (IC 50 of 9 lm) was found to be  10–15 times less potent than native CT. We then compared the rate of hydrolysis of the indi- vidual CT-A and CT-B subunits by cathepsin D at pH 4, followed by nonreducing SDS ⁄ PAGE and Coo- massie Blue staining (Fig. 7B). Pure cathepsin D was found to degrade the CT subunits in a manner similar to that observed for the hepatic endosomal fractions (Fig. 2A), i.e. CT-B subunit (t 1 ⁄ 2 < 15 min) > CT-A subunit (t 1 ⁄ 2  60 min). Proteolytic activation of CT induced by cathepsin D treatment CT cleavage induced by cathepsin D may be an essen- tial step for the cytotoxic activity of the toxin. Conse- quently, we examined whether under conditions where CT was partially processed by cathepsin D, we would observe a corresponding change in the toxin cytotoxi- city (Fig. 8). By western blot analysis, optimal degra- dation of native CT was observed with a cathepsin D concentration of 40 UÆmL )1 Æmg )1 (Fig. 8A). After a 1 min incubation at this concentration, the 25-kDa proteolytic fragment of the CT-A subunit previously identified (Figs 2A and 4A) was detected. Another 22-kDa product, whose presence was observed in the control condition (lane -), was prominently detected in CT hydrolysates performed using 4 UÆmL )1 Æmg )1 cath- epsin D. Therefore, the 40 UÆmL )1 Æmg )1 concentration was used to follow the subsequent proteolytic activa- tion of the toxin. CT was first partially processed by cathepsin D at pH 4–7 using the above enzyme ⁄ sub- strate ratio and then incubated at neutral pH with microsomal membranes in the presence of [ 32 P]NAD (Fig. 8B, upper panel). A rapid ADP-ribosylation of microsomal Gsa was evident following proteolysis of CT at acidic pH (pH 4–5), especially after a 15 min digestion at pH 5. In vitro digestion of CT at pH 7, a pH at which cathepsin D activity is nonexistent [12], did not reveal any Gsa labeling even after 60 min of proteolysis. Comparably, no detectable 32 P-labeling of Gsa was observed after treatment of CT under acidic (pH 4.5 and 5.5) or neutral (pH 7.5) conditions (Fig. 8B, middle panel). However, in vitro reduction of native CT or the CT-A subunit at 37 °C and pH 7 Fig. 5. Effect of immunodepletion of cathep- sins on endosomal CT-degrading activity. (A) ENs fractions were immunodepleted of act- ive cathepsin D (a-CD) or cathepsin B (a-CB) using their respective polyclonal antibodies which had been precoated onto protein G- Sepharose beads. Following centrifugation, the resultant supernatants were incubated with 1 l M native CT in 175 mM citrate ⁄ phosphate buffer pH 4 for 120 min at 37 °C, and then analyzed by RP-HPLC. The rate of proteolysis of CT-A and CT-B subunits was determined by following the disappear- ance of the peak area corresponding to the parent peptides. (B) ENs were immunode- pleted of active cathepsin D using polyclonal antibodies directed against human (M8147) or mouse (R291) cathepsin D, and the resultant supernatants were tested for their ability to degrade native CT at pH 4 as described for panel A. The proteolytic reac- tion was stopped with acetic acid (15%) and the samples were analyzed by RP-HPLC. Shown are absorbance profiles at 214 nm. Endosomal proteolysis of cholera toxin by cathepsin D C. Merlen et al. 4390 FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS with dithiothreitol (0.2 m final concentration) revealed [ 32 P]ADP-ribose incorporation into microsomal Gsa (Fig. 8B, lower panel), as reported previously [15]. The possibility that the CT-A subunit was artefactually reduced during the treatment with cathepsin D in vitro was ruled out by the absence of production of the A1-peptide when native CT or CT-A subunit was incu- bated with cathepsin D in the presence of pepstatin-A at both pH 4 and 7 (results not shown). Discussion As hepatic parenchyma accounts for a large fraction of the total CT-binding sites in the body, subcellular frac- tions from rat liver have been used to characterize the initial step in the interaction of CT with cells [6,16,17]. The time-dependent internalization of CT in hepatic cells was later assessed biochemically using isolated he- patocytes [18] and morphologically using the rat liver cell line KLTRYPV [19], a mouse hepatocyte cell line [20] and primary culture hepatocytes [21]. In agreement with these studies, we have found that in vivo both CT-A and CT-B subunits undergo stoichiometric endocytosis in rat liver and hepatoma HepG2 cells more rapidly and to a greater extent than in intestinal Caco-2 cells. This may reflect, in part, the greater binding capacity of hepatocytes and hepatoma cells [6,16,17], the lower dis- sociation constant of the hepatocyte–toxin interaction [18] and ⁄ or the higher content of ganglioside GM1 in hepatocytes as compared to other cell types. Previously, subcellular fractionation techniques used to assess the in vivo localization of radiolabeled 125 I- labelled CT taken up by rat liver have shown that both radioactive A- and B-subunits sequentially associate with the plasma membrane, endosomes and lysosomes, and that a very low level of A1 peptide occurred in the endosomal compartment (< 5% of total endosomal radioactivity recovered at 30 min postinjection) [6]. Unfortunately, in these studies, the nature of the endo- somal alterations of the CT-A subunit was based on (a) comparison of the electrophoretic mobility of the endosomal radioactive CT-A subunit to that of the chemically reduced 125 I-labelled CT-A subunit; and (b) TCA-precipitation which greatly underestimates CT-A subunit degradation and reduction [6,18]. Moreover, some of the intermediates might not have been radio- labeled and therefore would have escaped detection. Finally, the nature of the relevant enzymatic activity (protease and ⁄ or reductase) was not investigated. In addition, the fraction of hepatocyte-associated 125 I- labelled CT that was converted into A1 peptide was < 4% after 60 min, suggesting that the endo- somal reductive pathway may represent a minor meta- bolic fate for internalized CT within the endosomal apparatus [18]. Alternatively, the radioactive iodine on the CT-A subunit might have had a detrimental effect A B C Fig. 6. Partial colocalization of internalized CT and cathepsin D in hepatoma HepG2 and intestinal Caco-2 cells.HepG2 (A and B) and Caco-2 cells (C) were treated with 1 l M native CT for 30 min (B) or 1 lM CTB-FITC for 15 min (A and C), fixed with paraformal- dehyde and permeabilized with Triton X-100 prior to staining. In panels A and C, HepG2 and Caco-2 cells were treated with poly- clonal anti-(cathepsin D) Ig R291. In Panel B, HepG2 cells were treated with both poly- clonal anti-(cathepsin D) Ig R291 and anti- CTB monoclonal antibody D15-8. The mono- clonal antibody D15-8 directed against the B-subunit was concluded to be highly speci- fic as it recognized only the B-subunit by western blot analysis (results not shown). CT is shown in green and cathepsin D is shown in red. Merged images on the right indicate the extent of colocalization (yellow). Scale bar, 7.6 lm (A and B) or 8.5 lm(C). Fluorescent images were captured at two emission wavelengths (488 and 543 nm). C. Merlen et al. Endosomal proteolysis of cholera toxin by cathepsin D FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS 4391 on the endosomal proteolytic and reductional systems by affecting the degradation and ⁄ or reduction rates. Although unable to document CT reduction in endo- cytic structures, our in vivo and in vitro endosome studies clearly show that, under conditions ensuring acidification of these structures, internalized unlabeled CT-A subunit was progressively processed into degra- dative fragments. However, as a possible consequence of diffusion of degradative products out of the endo- somal apparatus and ⁄ or transfer of CT metabolites to lysosomes, a net detection of CT proteolytic fragments was only observed in our in vitro assays. Exploiting the fact that two substrates competing for the same enzyme will inhibit each other [14], we have defined the affinity of native CT and its individ- ual subunits for the cathepsin D protease. On the basis of IC 50 values, the CT-A and CT-B subunits and HI share a common binding site on the cathepsin D enzyme. Competition studies revealed that native CT and CT-B subunit displayed nearly equivalent affinity for cathepsin D (IC 50 of 0.6–0.9 lm). However, the CT-A subunit and the cathepsin D substrate HI [12] were found to be 10–15 times less potent. These competition studies correlated with results obtained from degradation studies using endosomal fractions or pure bovine cathepsin D in which the rate of peptide hydrolysis was found to be low with the A-subunit and high with the B-subunit. Interestingly, insulin elicits inhibition of CT-stimulated adenylate cyclase activity in both hepatocytes and the P9 immortalized hepatocyte cell line [22]. Attenuating effects on CT action may well originate from the ability of both pep- tides to accumulate into hepatic endocytic vesicles and to interact with endosomal cathepsin D (this study and [12]). Comparably, it has been reported that the CT-B subunit alters the progression of exogenous antigens along the endocytic processing pathway, and prevents or delays efficient epitope presentation and T-cell sti- mulation [23]. Studies have suggested the potential role of cathepsin D in conjunction with cathepsins S and L in degrading endocytosed antigens and the invariant chain within antigen-presenting cells [24]. CT may have the ability to alter the immune response by reducing cathepsin D activity and its subsequent processing of antigen in antigen-presenting cells [23]. In either case, it is clear that the endosomal cathepsin D activity exhibits a high specificity that limits further proteolysis of the bioactive CT-degrading fragment. The present studies indicate that both CT-A and CT-B subunits are high-affinity substrates for the endosomal protease cathepsin D. We confirmed our earlier biochemical and morphological studies that had localized active cathepsin D to endosomal subcellular fractions of rat liver [11,12] and early endocytic EEA1- positive vesicles of rat hepatocytes [12]. Thus, we have previously reported that internalized glucagon [11] and insulin [12] are partially processed within hepatic endo- somes by the endopeptidase activity of cathepsin D. The participation of cathepsin D in the endosomal proteolysis of mannose-BSA and parathyroid hor- mone-(1)84), in macrophages [25,26], invariant chain and endocytosed antigens in antigen-presenting cells [24] and b-amyloid precursor protein in astrocytoma cells [27] has been comparably demonstrated. Import- antly, the role of cathepsin D in conjunction with cath- epsin B in the endosomal processing, membrane translocation and cytotoxicity of ricin A chain has also been established [28]. CT-A – 515306090 – 5 15 30 60 (min of incubation) CT-A (28-kDa) CT-B (12-kDa) CT-B Fig. 7. Affinity-binding and degradation of native CT, CT-A subunit and CT-B subunit by cathepsin D. (A) Competition of native CT, CT- A subunit, CT-B subunit and HI for the degradation of 125 I-labelled Tyr A14 -HI by cathepsin D. Bovine cathepsin D (0.025 UÆmL )1 )was incubated with 125 I-labelled Tyr A14 -HI (75 fmol) for 20 min at 37 °C in 0.1 M citrate-phosphate buffer pH 4 with the indicated concentra- tions of unlabeled peptides. The amount of degraded radiolabeled insulin was determined by precipitation with TCA. Results are the mean of three separate experiments and are expressed as a per- centage of degradation observed in the absence of added unlabeled peptides. (B) CT-A or -B subunit (7 lg) was incubated with cathep- sin D (40 UÆmL )1 Æmg )1 )at37°C in citrate-phosphate buffer pH 4 for the indicated times. The incubation mixtures were then ana- lyzed by nonreducing SDS ⁄ PAGE followed by Coomassie Brilliant Blue staining. Arrows indicate the mobility of intact CT-A (28 kDa) and CT-B (12 kDa) subunits. Endosomal proteolysis of cholera toxin by cathepsin D C. Merlen et al. 4392 FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS The endosomal CT-degrading activity described here differs from the neutral Ca 2+ -dependent furin enzyme which participates in the proteolytic activation of other bacterial toxins [29] as follows: (a) furin recognition sequences (-Arg-X-Lys ⁄ Arg-Arg- or -Lys ⁄ Arg-X-X-X- Lys ⁄ Arg-Arg-) are not present in the CT molecule [30]; (b) no degradation products were observed when pure active furin was incubated with CT (results not shown); and (c) the acidic CT-degrading activity was not inhibited by the metal-chelating agent EDTA, whereas Ca 2+ and a neutral pH were strictly required for furin activity [31]. Correlations between CT action, endocytosis and processing were first made in cell fractionation studies using rat hepatocytes intoxicated with CT in vivo [6,7]. In these experiments, the endosomal processing of the A-subunit was closely associated with activation of adenylate cyclase and CT-induced activation of adeny- late cyclase appeared to depend on endosome acidifica- tion [7,18]. Thus, the acidotropic drug chloroquine reduced adenylate cyclase activation by CT and length- ened the lag phase both in vivo [7] and in vitro [18]. In addition intracellular CT, after a 30–60 min lag phase, was localized within the ER in murine hepatocyte BNL CL.2 cells [20] and other various cell types [32] using electron microscopy, fluorescence assay and sub- cellular fractionation. That CT followed a retrograde pathway into the Golgi and then the ER, with subse- quent ER processing participating in CT cytotoxic action, was provided by four independent lines of evidence: (a) inactivating mutations or removing the KDEL motif attenuated the efficiency of toxin action greater than 10-fold in polarized intestinal T84 cells [33]; (b) inhibition of CT action correlated with brefeldin A- induced disruption of Golgi structure and function [34]; (c) reduction of the A-subunit was shown within the ER lumen in human intestinal Caco-2 cells and depended on catalysis by PDI [35]; (d) interaction between the A1-peptide and the protein translocation channel of the ER was shown, followed by transport of the A1-peptide through the Sec 61p translocon into the cytosol [36]. In cell types other than hepatocytes, CT-induced cellular responses were brefeldin A-sensitive but independent of organelle acidification (chloroquine-insensitive) [37], suggesting that the dual-compartmental activation of intracellular CT may be cell-type specific. The discrep- ancies between these data sets may also be due to the experimental approaches: an in vivo model used for the A B Fig. 8. Effect of cathepsin D treatment on CT-catalyzed ADP-ribosylation of microsomal Gsa. (A) Native CT (7 lg) was incubated in vitro with cathepsin D (40 or 4 UÆmL )1 Æmg )1 )at37°C in citrate-phosphate buffer pH 4 for the indicated times. The incubation mixtures were then ana- lyzed by western blotting using the polyclonal anti-CT Ig. Molecular mass markers are indicated to the left. Arrows indicate the mobility of the intact A ( 28 kDa) and B subunits ( 12 kDa). (B) Native CT and ⁄ or CT-A were digested in vitro with cathepsin D (40 UÆmL )1 Æmg )1 )at 37 °C in citrate-phosphate buffer pH 4–7 for the indicated times or incubated at 37 °C in citrate-phosphate buffer pH 4.5–7.5 with or without dithiothreitol (0.2 M) for 30 min. The treated CT (5 lg) was then incubated for 45 min at 30 °C with microsomal proteins ( 50 lg) in sodium phosphate buffer pH 7.2 in the presence of 0.52 l M [ 32 P]NAD. Samples were then subjected to SDS ⁄ PAGE and analyzed by autoradio- graphy. Dithiothreitol and cathepsin D samples were exposed to X-ray films at )80 °C for 1 or 5 days, respectively. The arrow indicates the mobility of 32 P-labeled Gsa ( 45 kDa). C. Merlen et al. Endosomal proteolysis of cholera toxin by cathepsin D FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS 4393 studies on hepatocytes [6,7] and an in vitro model for the studies on intestinal and other cell types [34–37]. On the other hand, CT is internalized into cells by both clathrin-dependent and -independent endocytosis and these pathways may be differentially affected by acido- tropic agents and carboxylic ionophores. Previous studies have shown that CT administration to rats markedly and significantly increased rat liver endosome acidification [38]. The more acidic pH of these endocytic vesicles might facilitate (a) the inter- action of internalized CT with endosomal cathepsin D which displays an optimum activity at pH 4; and (b) the translocation of the A-subunit (or A-subunit frag- ment) across the endosomal membrane, which requires a low pH [18]. The subcellular site where endosomal activated CT ADP-ribosylates Gsa protein and activates adenylate cyclase remains to be clarified. Several lines of evidence now indicate that trimeric G proteins and adenylate cyclase are located in the endosomal compartment of various cells [39–41]. Previous studies report the pres- ence of the a-subunit of Gs protein [40], a fluoride-sen- sitive adenylate cyclase activity [7,42,43] and adenylate cyclase VI [40] in rat liver endosomal fractions. These observations are also in agreement with related studies on endosomal fusion where Gsa appears to be implica- ted as fusion between endosomes was abolished using CT in the presence of NAD, an antibody against the C-terminus of Gsa or synthetic peptides that preferen- tially activate Gsa [44]. Finally, activation of hepatic adenylate cyclase after CT injection into rats occurs first in endosomal fractions and, later, in plasma mem- brane fractions, suggestive of sequential activation of adenylate cyclase in these compartments [7]. In summary, we have characterized in vivo and in vitro the metabolic fate of internalized CT within the endosomal apparatus of rat liver. We have found that proteolytic degradation of CT was mediated by endosomal cathepsin D which assisted in the release of CT-A fragment(s) that are active towards the Gsa protein. Studies are currently underway to elucidate the sites of cleavage of CT by cathepsin D and to determine the endosomal degradative CT-A frag- ment(s) responsible for the endosomal ADP-ribosyla- tion of the Gsa substrate. Experimental procedures Peptides, ligand radioiodination, antibodies, protein determination and materials CT-A and -B subunits, native CT, CT-B subunit-FITC, HI and bovine cathepsin D (EC 3.4.23.5), 15 UÆmg )1 , were purchased from Sigma (St Louis, MO, USA). HI was radioiodinated by the lactoperoxidase method and purified by RP-HPLC to specific activities of 150– 300 lCiÆlg )1 as previously described [45]. Rabbit polyclo- nal anti-(CT C3062) was from Sigma. Mouse monoclonal antibody directed against CT-B subunit (anti-CTB D15-8) was a gift from F. Nato (Institut Pasteur, Paris, France). Rabbit anti-(mouse cathepsin D R291) Ig [11,12], sheep anti-(human cathepsin D M8147) [12] and rabbit anti-(rat cathepsin B 7183) Ig [13] were obtained from J.S. Mort (Shriners Hospital for Crippled Children, Montreal, Quebec) and used to immune deplete samples of native mature enzymes as described previously [11–13]. HRP- conjugated goat anti-(rabbit IgG) Ig was from Bio-Rad (Hercules, CA, USA). The protein content of isolated fractions was determined by the method of Lowry et al. [46]. Nitrocellulose membranes and Enhanced Chemi- Luminescence (ECL) detection kit were from Amersham. Protein G-Sepharose was from Pharmacia (Peapack, NJ, USA). Pepstatin-A, E-64 and PMSF were from Sigma. HPLC grade acetonitrile and trifluoroacetic acid (TFA) were obtained from Baker Chemical Co. (Phillipsburg, NJ, USA). All other chemicals were obtained from com- mercial sources and were of reagent grade. Animals and injections In vivo procedures were approved by the INSERM committee for use and care of experimental animals. Male Sprague-Dawley rats, body weight 180–200 g, were obtained from Charles River France (St. Aubin Les Elbeufs, France) and were fasted for 18 h prior to being killed. Native CT or CT-B subunit (50 lg per 100 g body weight) in 0.4 mL of 0.15 m NaCl was injected within 5 s into the penile vein under light anaesthesia with ether. Isolation of subcellular fractions from rat liver Subcellular fractionation was performed using established procedures [11–13,45]. Following injection of native CT or CT-B subunit, rats were killed and livers rapidly removed and minced in isotonic ice-cold homogenization buffer as previously described [11–13,45]. Microsomal (P) fraction was isolated by differential cen- trifugation as previously described [11–13,45]. The endo- somal (EN) fraction was isolated by discontinuous sucrose gradient centrifugation and collected at the 0.25–1.0 m sucrose interface [11–13,45]. The soluble endosomal extract (ENs) was isolated from the EN fraction by freeze ⁄ thawing in 5 mm Na-phosphate pH 7.4, and disrupted in the same hypotonic medium using a small Dounce homo- genizer (15 strokes with Type A pestle) followed by cen- trifugation at 150 000 g for 60 min as previously described [11–13,45]. Endosomal proteolysis of cholera toxin by cathepsin D C. Merlen et al. 4394 FEBS Journal 272 (2005) 4385–4397 ª 2005 FEBS [...]... cyclase by cholera toxin requires toxin internalization and processing in endosomes J Biol Chem 266, 12858–12865 8 Sandvig K & Van Deurs B (2002) Membrane traffic exploited by protein toxins Annu Rev Cell Dev Biol 18, 1–24 9 Majoul I, Ferrari D & Soling HD (1997) Reduction of ¨ protein disulfide bonds in an oxidizing environment The disulfide bridge of cholera toxin A-subunit is reduced in the endoplasmic reticulum... The proteolytic reaction was stopped by addition of nonreducing SDS ⁄ PAGE buffer (for western blot analysis and Coomassie Brilliant Blue staining) or ADP-ribosylation buffer (for ADP-ribosylation analysis) For the in vitro degradation of 125I-labelled TyrA14-HI by bovine cathepsin D, the radiolabeled HI (70 fmol) was incubated with 0.01 UÆmL)1 cathepsin D and various concentrations of unlabeled native... degradation by endosomal acidic insulinase Endocrinology 142, 276–289 15 Ribeiron-Neto F, Mattera R, Grenet D, Sekura RD, Birnbaumer L & Field JB (1987) Adenosine diphosphate ribosylation of G proteins by pertussis and cholera toxin in isolated membranes Different requirements for and effects of guanine nucleotides and Mg2+ Mol Endocrinol 1, 472–481 16 Cuatrecasas P (1973) Interaction of Vibrio cholerae... pretreated CT (5 lg) in an ADP-ribosylation buffer containing 0.54 lm [32P]NAD, 50 mm sodium phosphate buffer pH 7.2, 0.5 mm GTP, 1 mm ATP, 5 mm MgCl2 and 10 mm thymidine for 45 min at 30 °C The reaction was stopped by the addition of Laemmli sample buffer [47] followed by SDS ⁄ PAGE and autoradiography Immunodepletion studies ENs was immunodepleted of active cathepsin B or cathepsin D prior to the digestion... step by incubating ENs (0.15 mgÆmL)1) with antibodies coated onto protein GSepharose beads for 16 h at 4 °C in 800 lL of 20 mm sodium phosphate buffer (pH 7) The fractions were then centrifuged for 5 min at 10 000 g, and the resultant immunodepleted supernatants were used in the toxin degradation assay The reaction was terminated by the addition of 15% (v ⁄ v) acetic acid and immediately analyzed by. .. region of the B chain by cathepsin D J Biol Chem 277, 9437–9446 ´ 13 Authier F, Metioui M, Bell AW & Mort JS (1999) Negative regulation of epidermal growth factor signaling by selective proteolytic mechanisms in the endosome mediated by cathepsin B J Biol Chem 274, 33723–33731 14 Authier F, Danielsen GM, Kouach M, Briand G & Chauvet G (2001) Identification of insulin domains important for binding to and degradation... 37 °C The amount of radiolabeled HI-degraded was assayed by precipitation with 2 mL of ice-cold 10% (v ⁄ v) TCA for 2 h at 4 °C The samples were then centrifuged at 10 000 g for 20 min at 4 °C, and the supernatants and pellets evaluated for their radioactive content using a Packard c-counter (Perkin Elmer, Wellesley, MA, USA) CT-catalyzed ADP-ribosylation Native CT was first digested by incubation at... Cholesterol depletion by methyl-beta-cyclodextrin blocks cholera toxin transport from endosomes to the Golgi apparatus in hippocampal neurons J Neurochem 78, 991–999 4396 C Merlen et al 6 Janicot M & Desbuquois B (1987) Fate of injected 125Ilabeled cholera toxin taken up by rat liver in vivo Eur J Biochem 163, 433–442 7 Janicot M, Fouque F & Desbuquois B (1991) Activation of rat liver adenylate cyclase by cholera. .. Herzog V (2000) Cholera toxin is exported from microsomes by the Sec61p complex J Cell Biol 148, 1203–1212 11 Authier F, Mort JS, Bell AW, Posner BI & Bergeron JJM (1995) Proteolysis of glucagon within hepatic endosomes by membrane-associated cathepsins B and D J Biol Chem 270, 15798–15807 ´ 12 Authier F, Metioui M, Fabrega S, Kouach M & Briand G (2002) Endosomal proteolysis of internalized insulin at... CT C3062 (diluted 1 : 60000)] in the above buffer for 16 h at 4 °C The blots were then washed 3 times with 0.5% (w ⁄ v) skimmed milk in 10 mm Tris ⁄ HCl pH 7.5, 300 mm NaCl and 0.05% (v ⁄ v) Tween-20 over a period of 1 h at room temperature The bound Ig was detected using HRP-conjugated goat anti-(rabbit IgG) Ig In vitro proteolysis of CT peptides by hepatic endosomes and cathepsin D ENs ( 1 lg) or . acid protease cathepsin D in the endosomal degradation of internalized CT. We show that hydrolysates of CT generated in vitro by pure cathepsin D displayed ADP-ribosyltransferase activity towards. ability of hepatic endosomes to degrade CT (Fig. 2). The luminal and membrane- bound distribution of endosomal CT-degrading activity was assessed by western blot analysis of CT digestions performed. Proteolytic activation of internalized cholera toxin within hepatic endosomes by cathepsin D Cle ´ mence Merlen*, Domitille Fayol-Messaoudi*, Sylvie Fabrega, Tatiana El Hage, Alain Servin and