Ricin is a potent plant cytotoxin composed of an A-chain [RTA (ricin A-chain)] connected by a disulfide bond to a cell binding lectin B-chain [RTB (ricin B-chain)]. After endocytic uptake, the toxin is transported retrogradely to the ER (endoplasmic reticulum) from where enzymatically active RTA is translocated to the cytosol. This transport is promoted by the EDEM1 (ER degradation-enhancing α-mannosidase I-like protein 1), which is also responsible for directing aberrant proteins for ERAD (ER-associated protein degradation). RTA contains a 12-residue hydrophobic C-terminal region that becomes exposed after reduction of ricin in the ER. This region, especially Pro250, plays a crucial role in ricin cytotoxicity. In the present study, we introduced a point mutation [P250A (substitution of Pro250 with alanine)] in the hydrophobic region of RTA to study the intracellular transport of the modified toxin. The introduced mutation alters the secondary structure of RTA into a more helical structure. Mutation P250A increases endosomal–lysosomal degradation of the toxin, as well as reducing its transport from the ER to the cytosol. Transport of modified RTA to the cytosol, in contrast to wild-type RTA, appears to be EDEM1-independent. Importantly, the interaction between EDEM1 and RTAP250A is reduced. This is the first reported evidence that EDEM1 protein recognition might be determined by the structure of the ERAD substrate.
- endoplasmic reticulum degradation-enhancing α-mannosidase I-like protein 1 (EDEM1)
- endoplasmic reticulum (ER)
- endosomal–lysosomal degradation
- P250A mutation
- ricin A-chain (RTA)
Ricin, the most toxic member of the RIPs (ribosome-inactivating proteins), accumulates to high levels in the endosperm of Ricinus communis seeds. Cell trafficking of this toxin has been extensively studied due to the variety of medical applications in which it can be used. As an example, ricin has been employed in the construction of ITs (immunotoxins) that show specific anti-cancer and anti-AIDS activities in vitro and in vivo [1–4]. Ricin is a heterodimeric glycoprotein composed of an RNA-specific N-glycosidase A-chain [RTA (ricin A-chain)] connected by a disulfide bond to a cell binding lectin B-chain [RTB (ricin B-chain)]. RTA inhibits protein synthesis by enzymatically depurinating a specific adenine residue at the sarcin–ricin loop of the 28S rRNA, thereby preventing the binding of elongation factors to the GTPase activation centre of the ribosome . RTB recognizes cell surface glycolipids or glycoproteins with β-1,4-linked galactose residues. On binding to the cell surface, the toxin is taken up by endocytosis and enters early endosomes. From here, the majority of the endocytosed toxin recycles back to the cell surface, starts to be degraded or proceeds to late endosomes and lysosomes where further degradation is conducted [6–8]. A minor fraction (~ 5%) of ricin is transported from early endosomes to the Golgi apparatus, and from there it is further transported retrogradely to the ER (endoplasmic reticulum). In the ER, the disulfide bond of the holotoxin is reduced to liberate RTA. Upon A- and B-chain dissociation, a 12-amino-acid (Val245 to Val256) hydrophobic region of the RTA, which is hidden in the holotoxin, becomes exposed. It has been demonstrated that P250A (substitution of Pro250 with alanine) in this region results in a dramatic decrease in RTAP250A cytotoxicity in Vero cells . It is possible that exposure of the RTA hydrophobic region in the ER triggers an interaction between RTA and membranes, ER chaperones or even ER translocons. Partial unfolding of RTA renders it competent to cross the ER membrane in a similar manner as misfolded ER proteins that are dispatched by proteasomal degradation via the ERAD (ER-associated protein degradation) pathway [10,11]. Retrotranslocation of RTA from the ER to the cytosol is believed to occur via the Sec61p translocon  with assistance of an ER chaperone, EDEM1 (ER degradation-enhancing α-mannosidase I-like protein 1) . In the cytosol RTA must refold into its biologically active conformation to inactivate the ribosomes. It has been demonstrated that Hsc70 (heat-shock cognate 70 stress protein) and Hsp90 (heat-shock protein 90) cytosolic chaperone machines are involved in RTA folding after retrotranslocation to the cytosol . There is evidence suggesting that RTA can partially disconnect from the ERAD pathway by virtue of its low lysine content .
ERAD is a component of the protein quality control system ensuring that misfolded ER proteins are recognized and targeted for degradation. EDEM1 accelerates ERAD of misfolded glycoproteins; its overexpression results in faster release of folding-incompetent proteins from the cnx (calnexin) cycle and earlier onset of their degradation, whereas EDEM1 down-regulation delays ERAD by prolonging folding attempts [16,17]. It has been demonstrated that EDEM1-dependent disposal of glycoprotein is regulated by accelerating substrate de-mannosylation by this protein. EDEM1 removes mannose residues from branches A and C of glycoproteins, thereby preventing their re-glucosylation and return of folding-incompetent proteins to the cnx chaperone system [18,19]. Moreover, it was shown that EDEM1 accelerates ERAD by inhibiting the formation of disulfide-bonded dimers  or covalent aggregates on release of terminally misfolded ERAD candidates from cnx . Non-glycan-mediated interaction of EDEM1 and protein substrates has also been reported for the following: the protein toxin ricin , NHK (null Hong Kong, mutant variant of α1-antitrypsin), the mutant variant of α1-AT (α1-antitrypsin)  and the mutant P23H rod opsin . Despite the suggestions that EDEM1 recognizes both glycans and putative misfolded regions of aberrant proteins, little is known about the general mechanisms of substrate recognition by EDEM1 and sorting to the ERAD pathway. This knowledge is crucial in understanding the quality control system operating in the ER, knowledge that may give a better understanding of a broad variety of human diseases caused by defective protein folding or trafficking.
In the present study, we produced ricin with a point mutation in the hydrophobic region of RTAP250A in order to investigate the effect of this mutation on vesicular transport of ricin, its retrotranslocation from the ER to the cytosol and the interaction between EDEM1 and RTAP250A. The P250A mutation alters the secondary structure of the toxin that results in a more helical structure of the protein. We have demonstrated that modified ricin is more extensively degraded in endosomes/lysosomes than the wild-type protein; moreover, we show here that a smaller fraction of P250A ricin is retrotranslocated from the ER to the cytosol when compared with its wild-type counterpart. Importantly, our study revealed that in contrast to wild-type ricin, retrotranslocation of the ricin P250A mutant to the cytosol is EDEM1-independent. Pull-down and co-immunoprecipitation experiments showed that the interaction between EDEM1 and RTAP250A was significantly decreased. The implications of these findings for medical applications involving ricin, as well as for our understanding of ERAD, are discussed.
RTB was obtained from Vector Labs (Burlingame, CA, U.S.A.), Hepes, α-lactose monohydrate, trypsin, proteinase K, pronase, bafilomycin A1, brefeldin A, lactacystin, pepstatin A, CA074 methyl ester and digitonin came from Sigma–Aldrich (St. Louis, MO, U.S.A.). [3H]Leucine was purchased from GE Healthcare (Princeton, NJ, U.S.A.), Na235SO4 came from Hartmann Analytic (Braunschweig, Germany), whereas Na-125I from DuPont (Brussels, Belgium). The mouse monoclonal anti-HA antibodies were obtained from Covance Research Products (Denver, CO, U.S.A.), rabbit anti-Ricinus communis-lectin antibodies were obtained from Sigma–Aldrich, whereas mouse monoclonal anti-RTA and sheep anti-TGN46 (TGN46 is trans-Golgi network protein, 46 kDa) antibodies were purchased from Serotec (Oxford, U.K.). The rabbit anti-His antibodies came from Santa Cruz Biotechnology (Santa Cruz, CA, U.S.A.). The mouse anti-cnx (BD Biosciences, Palo Alto, CA, U.S.A.), anti-crt (calreticulin) (BioSite, San Diego, CA, U.S.A.) and anti-α-tubulin (Sigma–Aldrich) were used in Western blots. The secondary anti-rabbit HRP (horseradish peroxidase) and anti-mouse HRP antibodies were obtained from Sigma–Aldrich, whereas anti-rabbit Alexa Fluor® 555 and anti-sheep Cy-2 were obtained from Jackson laboratories (Bar Harbour, MA, U.S.A.).
HEK-293 cells (human embryonic kidney cells) and African green monkey kidney (Vero) cells were grown under 5% CO2 in DMEM (Dulbecco's modified Eagle's medium; Invitrogen, Carlsbad, CA, U.S.A.) supplemented with 10% FBS (fetal bovine serum) (PAA, Pasching, Austria), 25 units/ml penicillin and 25 μg/ml streptomycin (all from Invitrogen) in a 37 °C incubator.
Cloning, mutagenesis, cDNA constructs and transfections
Ricin A was amplified from pRA (a gift from Professor Sjur Olsnes, Department of Biochemistry, Norwegian Radium Hospital, Oslo, Norway) . Detailed cloning and mutagenesis are described in the Supplementary online data (available at http://www.BiochemJ.org/bj/436/bj4360371add.htm)
cDNA encoding the mouse EDEM1 fused to a HA (haemagglutinin) tag in the pCMV-SPORT2 vector was a gift from Professor K. Nagata and Dr N. Hosokawa (Kyoto, Japan). For details on the cloning, refer to Hosokawa et al. .
HEK-293 cells were transiently transfected with FuGENE™ transfection reagent (Roche Diagnostics, Basel, Switzerland) according to the manufacturer's procedure or TurboFect (Fermatas, Vilnius, Lithuania).
Purification of RTA proteins and reconstitution with RTB
His-tagged RTA and His-tagged modified P250A RTA were expressed in Escherichia coli Rosetta cells (Merck) and purified using Ni-NTA (Ni2+-nitrilotriacetate) agarose beads (Qiagen, Germantown, MD, U.S.A.) according to the manufacturer's manual. The eluate was finally dialysed overnight in PBS.
RTA sulf-1 and mutant P250A RTA sulf-1 fused to MBP (maltose-binding protein) were applied to a column with amylose resin and purified as previously described . Free wild-type RTA and RTAP250A were cleaved off with Factor Xa (New England Biolabs, Ipswich, MA, U.S.A.). For further purification, wild-type RTA and RTAP250A proteins were applied on to a MonoS column (GE Healthcare) and purified using GE Pharmacia Acta Purifier (GE Healthcare). Then 25 mM phosphate buffer, pH 6.5, was used as the column equilibrating buffer and the wash buffer, and proteins were eluted with 0–500 mM NaCl gradient, and the fractions containing wild-type RTA or RTAP250A were identified by Coomassie Blue staining of SDS/PAGE gels. Purified wild-type RTA and mutant P250A were mixed with the RTB and dialysed extensively against PBS to remove reducing agents.
Protease digestion assay
RTA or RTAP250A (500 ng) were incubated with increasing concentrations of trypsin (0–100 μg/ml) in NaCl/Pi (137 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4 and 1.8 mM KH2PO4, pH 7.4) at 37 °C for 15 min and then visualized by SDS/PAGE and Coomassie Blue staining. For proteinase K digestion, 500 ng of RTA or RTAP250A were incubated with the protease at a final concentration of 0.25 or 0.5 mg/ml in a buffer containing 20 mM Tris/HCl, pH 8.0, and 100 mM NaCl, at 0 °C for 40 min. Proteins were precipitated with cold 10% (w/v) trichloroacetic acid, washed with 100% acetone and visualized by SDS/PAGE and Coomassie Blue staining. Pronase digestion was performed at 40 °C for 20 min in a buffer containing 0.1 M Tris/HCl, pH 5.5 or 6.5, and 0.5% SDS with increasing concentrations of the protease (0–1.5 mg/ml); inactivation of the protease was carried out at 80 °C for 10 min. Proteins were visualized by SDS/PAGE and Coomassie Blue staining.
Interchain disulfide bond stability
An investigation of disulfide bond stability was performed according to the previously published protocol .
Far-UV CD was measured on Jasco J-815 spectrapolarimeter (Jasco, Tokyo, Japan). Experiments were performed in 25 mM phosphate buffer, pH 6.5, using a 1-mm-pathlength cuvette. The concentration of the protein solutions was 30 μg/ml. Spectra recorded from 190 to 240 nm with a 1-nm step size were averaged from three accumulations and were corrected against the buffer. Secondary structure analysis of RTA and RTAP250A were performed by using DichroWeb  with analysis programs SELCON3 and CDSSTR. Unfolded RTA P250A was prepared by incubation of RTAP250A (30 μg/ml) at 65 °C for 20 min in 25 mM phosphate buffer (pH 6.5).
Measurements of protein synthesis
HEK-293 cells or Vero cells were washed in leucine-free Hepes-buffered medium and incubated with the same type of medium with different concentrations of wild-type or mutant P250A ricin for 3 or 12 h. Optionally, the cells were preincubated with lactacystin for 30 min before toxins were added. The concentration of lactacystin is indicated in the figure legend (Figure 2). The cells were then incubated in leucine-free medium supplemented with 1 μCi/ml [3H]leucine for 20 min or 2 h at 37 °C. Cells were extracted with 5% trichloroacetic acid for 20 min, followed by a wash (5 min) in 5% trichloroacetic acid and subsequently dissolved in 0.1 M KOH. The cell-associated radioactivity was measured. The results are expressed as the percentage of [3H]leucine incorporated into cells incubated without toxin. Deviations between duplicates did not vary by more than 10%.
Sulfation of wild-type ricin sulf-1 and mutant P250A and permeabilization of cells
Sulfation of wild-type ricin sulf-1 and mutant P250A and permeabilization of cells were performed according to the previously published protocol . Details on this method, as well as Western blot analysis of RTA, are described in the Supplementary online data.
Measurements of binding, endocytosis and degradation of wild-type ricin and mutant P250A
Ricin was 125I-labelled according to , to a specific activity of 3×104–5×104 cpm/ng. Methods used to measure the binding and endocytosis of wild-type and mutant P250A ricin are described in the Supplementary online data.
For degradation measurements, cells were incubated with or without bafilomycin A1 (0.1 μM) for 30 min, then 125I-labelled wild-type ricin or mutant P250A was added and incubation was continued for additional 15 min. After this, the cells were washed with Hepes medium containing 0.1 M lactose: 15 min incubation and three rapid washes. To determine the degradation of the toxin, incubation was continued for the next 3 h with or without bafilomycin A1 (0.1 μM). Degradation of wild-type ricin and mutant P250A was measured as the amount of radioactivity that could not be precipitated by trichloroacetic acid. In a second experiment, cells were treated either with bafilomycin A1, pepstatin A, CA074 methyl ester or a combination of pepstatin A and CA074 methyl ester, for 30 min, and then incubated with unlabelled wild-type ricin or mutant P250A for 3 h. To determine the total amount of ricin remaining in the cells after degradation, Western blot with anti-RTA antibodies was performed. Concentrations of inhibitors are indicated in the figure legend (Figure 5).
Confocal fluorescence microscopy
HEK-293 cells were grown on coverslips and incubated with wild-type or P250A ricin for 30 min or 3 h. The cells were then incubated twice (5 min) with a 0.1 M lactose solution at 37 °C, then washed once with PBS and fixed in 3% (w/v) paraformaldehyde (Sigma–Aldrich). Cells were then permeabilized in 0.1% Triton X-100 and blocked in 5% FBS before labelling with rabbit anti-ricin together with sheep anti-TGN46 and treated with the appropriate secondary antibodies. DRAQ5 (Alexis Biochemicals, San Diego, CA, U.S.A.) was used to stain the nuclei. The cells were mounted in Mowiol (Molecular Probes, Eugene, OR, U.S.A.) and examined with a laser-scanning confocal microscope LSM 510 META (Carl Zeiss, Jena, Germany). Images were prepared with the LSM Image Browser software (Carl Zeiss) and analysed by the JaCoP plugin  in the ImageJ software. Mander's co-efficient was used for reporting co-localization between ricin and TGN46. Mander's co-efficient ranged from 0 to 1, corresponding to non-overlapping images and 100% co-localization between the two images respectively.
Measurements of the amount of wild-type ricin and mutant P250A recycled back to the cell surface
Cells were incubated with 125I-labelled wild-type ricin or mutant P250A (~100 ng/ml) at 37 °C for 80 min. Ricin associated with the cell surface was then removed by washing the cells four times with 0.1 M lactose in Hepes at 37 °C: one long, 15-min wash and three rapid washes. The amount of wild-type ricin and mutant P250A recycled back to the cell surface was measured after a 40 min incubation in the presence of 1 mM lactose to prevent rebinding of recycled ricin to the cell surface. The trichloroacetic acid precipitation and non-precipitable radioactivity in the last incubation medium was measured. The obtained results were compared with the amount of endocytosed ricin.
His-tag pull-down assay
HEK-293 cells (105/plate) were seeded in 6 cm plates and transfected with either EDEM1-HA cDNA or an empty vector. Three days post-transfection cells were lysed (lysis buffer: 50 mM Hepes, pH 7.5, 150 mM NaCl, 10% glycerol and 1% Triton X-100) in the presence of a protease inhibitor mixture (Roche Diagnostics) and sonicated (5 s, 20% output). His-tag fusion proteins, wild-type RTA or RTAP250A (0.5 μg), were incubated at 37 °C for 2 h with lysates from cells transfected with the indicated constructs and then incubated with Ni-NTA–agarose beads (Qiagen). Beads were washed three times with lysis buffer supplemented with 30 mM imidazole and resuspended into an SDS/PAGE sample buffer. Amounts of toxin-bound EDEM1–HA were detected after Western blot with anti-HA antibodies. Signal intensities of the bands were quantified using ImageQuant 5.0 software (GE Healthcare). Membranes were re-probed with anti-His or anti-RTA antibodies to confirm equal amounts of RTA-His and RTAP250A–His used in the experiments.
EDEM1-transfected cells or cells transfected with an empty vector were incubated with 80 ng/ml wild-type ricin or mutant P250A for 3 h, washed with 0.1 M lactose solution at 37 °C, followed by a wash with cold HBS (Hepes-buffered saline; 20 mM Hepes and 150 mM NaCl, pH 6.8) and lysed in a buffer containing 2% CHAPS in HBS (pH 6.8), 20 mM N-ethylmalemide and protease inhibitor mixture (Roche Diagnostics). The supernatants were centrifuged to remove cell debris and nuclei at 10000 g. The co-immunoprecipitation assay was performed as described previously . Signal intensities of the bands were quantified using ImageQuant 5.0 software (GE Healthcare). For the control of an equal amount of EDEM1–HA in the probes, the same membranes were re-probed with anti-HA antibodies.
All experiments were performed independently at least three times. Values of three or more parallel experiments were given as a means±S.D. A P-value of 0.05 or less was considered to be statistically significant and determined by the Student's t test or ANOVA tests.
Biophysical characteristics of mutant P250A RTA
To investigate the significance of the C-terminal hydrophobic region of the RTA on intracellular transport of the toxin, we produced a mutated form of RTA containing an alanine substitution at amino acid position 250 to replace the naturally occurring proline residue (referred to as P250A, see the Experimental section).
To characterize the correct folding and overall stability of RTAP250A, its sensitivity to proteases was compared with that of wild-type ricin. Digestion patterns for the ricin P250A mutant and wild-type counterpart were the same for both trypsin (Figure 1A) and proteinase K (results not shown) confirming equal in vitro stability of mutant P250A and the wild-type protein. The secondary structure of RTAP250A was examined by CD and compared with the spectrum for unmodified RTA. The CD spectrum of RTAP250A indicates a higher amount of α-helices in comparison with wild-type RTA (Figure 1B). This increase in the amount of α-helical structures is concomitant with a decrease in β-sheet structures (DichroWeb analysis ; see the Experimental section). Wild-type RTA achieves accuracies of 0.19 for helices, 0.32 for β-sheet, 0.21 for turns and 0.21 for unordered structures, whereas RTAP250A has 0.28 for helices, 0.26 for β-sheet, 0.20 for turns and 0.26 for unordered structures. Normalized root mean square deviation for the measurements was <0.05, R<0.1. Additional CD spectrum for unfolded RTAP250A is also shown (Figure 1B). Thus mutation in the hydrophobic region of RTA slightly changes the secondary structure of RTA by introducing additional α-helical structures.
The obtained modified RTA was re-associated with RTB to form a holotoxin (RTAP250A:RTB) (see Experimental section). To ensure that the introduced mutation was not affecting the reduction of the disulfide bond, either by increasing or decreasing the possibility of reductive cleavage, mutant P250A and wild-type holotoxin were incubated with increasing amounts of DTT (dithiothreitiol). The results showed no significant difference in the RTAP250A:RTB reducibility in comparison with wild-type holotoxin (Figure 1C).
Ricin P250A mutant reveals decreased cytotoxicity in Vero and HEK-293 cells
To investigate whether a mutation in the hydrophobic region of RTA influences its cytotoxicity both in Vero and HEK-293 cells, cells were incubated with increasing concentrations of wild-type or modified ricin for 12 h (Figures 2A and 2B) or for 3 h (Figure 2C and results not shown for Vero cells) and the protein synthesis was then measured. Results show that the P250A mutation reduces the cytotoxicity of ricin 9-fold (IC50) in both cell lines regardless of the incubation time with the toxin. These results reveal a decreased cytotoxic effect of the ricin P250A mutant; moreover, the observed effect is not limited to only a single cell line. A lower cytotoxic effect of the modified P250A toxin was also reported previously in Vero cells incubated for 12 h with ricin .
There was a possibility that the observed protection against ricin containing the P250A mutation might be due to the increased proteasomal degradation of the ricin P250A mutant after retrotranslocation to the cytosol. It is known that, despite low lysine content, ricin is partially degraded by proteasomes and this degradation can be inhibited by lactacystin . In the presence of this proteasomal inhibitor cells were sensitized to both wild-type ricin and mutant P250A (Figure 2D). However, even in the presence of lactacystin, the cells were still 9-fold less sensitive to the P250A mutant than to the wild-type toxin. Thus these results suggest that the lower cytotoxicity of ricin with a mutation in the hydrophobic region of RTA was not due to an increased degradation of this toxin after retrotranslocation to the cytosol.
Mutation in the hydrophobic region of RTA does not influence the cellular reductive release of RTA from the holotoxin, but affects the total amount of the holotoxin P250A mutant in the cell
As described above a mutation changing proline into alanine in the C-terminus of the RTA does not affect the reduction of holotoxin in vitro. However, we could not exclude the possibility that the release of the RTAP250A mutant from the holotoxin in the ER was affected by this mutation. To examine the degree of RTA release, cells were incubated with wild-type holotoxin or mutant P250A for 3 h. RTA and the intact toxin were then immunoprecipitated from the cell lysate, separated under non-reducing conditions and analysed with anti-RTA antibodies (Figure 3A). There was no significant effect of the mutation in the hydrophobic region of RTA on the fraction of A-chain released from the holotoxin (Figures 3A and 3B). Thus the mutation substituting proline with alanine at 250 residue of RTA does not have any effect on the in vivo reduction of the disulfide bond connecting the A- and B-chain of the toxin.
Interestingly, there was approx. 3-fold less holotoxin P250A mutant and released A-chain compared with the amount of wild-type toxin present in the cell (Figures 3A and 3C). To further investigate the reason for the observed intracellular decrease in RTAP250:RTB, the amounts of wild-type and P250A ricin sulfated in the Golgi complex were examined. For this purpose, wild-type ricin sulf-1 and mutant P250A , a modified ricin molecule containing a sulfation site in the A-chain, were used. When cells are incubated with 35SO42−, the A-chain becomes radioactively labelled due to the sulfotransferase in the TGN, and the fate of the 35SO42−-labelled ricin molecule can be studied. As shown in Figure 3(D), a decrease in the amount of the sulfated ricin holotoxin P250A mutant was also observed. This decrease was similar, approx. 3-fold (results not shown), to the observed reduction in the amount of unlabelled P250A ricin (Figures 3A and 3C). One possible explanation for the reduced amount of sulfated P250A holotoxin in the Golgi complex could be that less P250A ricin is transported to the Golgi complex in comparison with wild-type ricin. Another possibility is that the holotoxin P250A mutant was less efficiently sulfated than wild-type ricin in the TGN. However, since the total amount of the ricin P250A mutant present in the cell was decreased compared with wild-type toxin (Figures 3A and 3C) and the level of total cellular protein sulfation, both in cells incubated with wild-type and P250A ricin, was equal (results not shown); these results strongly suggest that less mutant than wild-type toxin is transported to the Golgi complex.
To further confirm this observation, we investigated by confocal microscopy the localization of wild-type or P250A ricin together with TGN46, a marker for the TGN. P250A ricin was found to co-localize with TGN46 to a much lower extent than the wild-type protein (Figure 4). This was observed after 30 min incubation with the toxin as well as after 3 h. The microscopic data clearly support our biochemical analysis.
Mutation P250A of RTA does not influence the binding and endocytosis of the modified holotoxin, but affects its endosomal–lysosomal degradation
To further investigate the reason for the reduced intracellular amount of the holotoxin P250A mutant, measurements of binding and endocytosis of RTAP250A:RTB were performed. The total binding of wild-type holotoxin and mutant P250A to the cell surface was studied by measurement of total 125I-labelled wild-type and P250A ricin bound to the cell surface (Supplementary Figures S1A and S1B http://www.BiochemJ.org/bj/436/bj4360371add.htm) and by measurement of binding of unlabelled toxins using anti-RTA antibodies and visualization by Western-blot analysis (Supplementary Figures S1C and S1D). In both cases, there was no difference in the amount of cell surface bound P250A holotoxin in comparison with the wild-type protein (Supplementary Figures S1A–S1D). These results suggest that the point mutation in the hydrophobic region of RTA (P250A) does not affect the binding of RTAP250A:RTB to the cell surface. The amounts of cell surface bound wild-type and P250A holotoxin were lower at 37 °C (Supplementary Figure S1B) than at 0 °C (Supplementary Figure S1A), in agreement with previous findings for HeLa cells . However, temperature did not influence the ratio of cell surface-bound modified holotoxin and wild-type protein (Figures S1A and S1B).
To test whether mutation in the hydrophobic region of RTA affects endocytosis of RTAP250A:RTB, amounts of lactose-resistant wild-type holotoxin and mutant P250A were analysed. Supplementary Figure S2(A) (available at http://www.BiochemJ.org/bj/436/bj4360371add.htm) shows that there are no significant differences between the uptake of 125I-labelled P250A holotoxin and 125I-labelled wild-type ricin. Furthermore, endocytosis of both wild-type and modified holotoxin was equally strongly reduced when ATP production was blocked by adding 2-deoxyglucose and sodium azide (Figures S2B and S2C). Together, these results show that the mutation in the hydrophobic region of RTA does not affect endocytosis of the holotoxin P250A mutant.
To investigate whether endosomal–lysosomal degradation of RTAP250A:RTB is changed in comparison with wild-type protein, 125I -labelled ricin degradation was measured. Figure 5(A) shows that mutation in the hydrophobic region of RTA increases degradation of the holotoxin P250A mutant more than 2-fold when compared with the wild-type protein. Bafilomycin A1, an inhibitor of the vacuolar H+-ATPase [29,30] decreases degradation of both unmodified ricin and RTAP250A:RTB in HEK-293 cells. This is in agreement with previous observations, showing that ricin degradation takes place in low-pH compartments in HeLa cells . Importantly, the level of degradation of wild-type and P250A holotoxin in the presence of bafilomycin A1 is almost equal, strongly indicating that the low-pH-dependent degradation machinery is at least partially responsible for the increased ricin P250A mutant degradation. These results were confirmed in the experiments where the total amount of wild-type and P250A ricin in cells treated with or without bafilomycin A1 were estimated by Western-blot analysis using anti-RTA antibodies. As shown in Figures 5(B) and 5(C), bafilomycin A1 increases intracellular amounts of wild-type ricin and mutant P250A approximately to the same level, strengthening the suggestion that the observed decrease in the amount of the ricin holotoxin P250A mutant in the cell (Figure 3) is caused by increased endosomal–lysosomal degradation of holotoxin containing a mutation in RTA.
It has been demonstrated that the aspartyl protease cathepsin D and cysteine protease cathepsin B both participate in ricin cleavage . To estimate if these proteases are responsible for increased P250A ricin degradation, specific inhibitors were employed. Pepstatin A is a potent inhibitor of aspartyl proteases including cathepsin D; CA074 methyl ester blocks cysteine cathepsins, especially cathepsin B. Figures 5(D)–5(I) show that both types of proteases are involved in ricin P250A degradation. However, it seems that cathepsin B is more important in this process (Figures 5F and 5G). In the presence of the CA074 methyl ester the level of P250A ricin is significantly increased. The combination of pepstatin A and CA074 methyl ester seems to increase the cellular level of P250A ricin even further (Figures 5H and 5I), but not much more than in cells treated only with the CA074 methyl ester (Figures 5F and 5G). Thus mainly cathepsin B is responsible for the endosomal–lysosomal degradation of the ricin holotoxin P250A mutant in the cell.
To study degradation time dependency, the amount of wild-type holotoxin and mutant P250A in the cells were analysed after 10, 30, 45 and 120 min incubations with the toxins (Figures 6A and 6B). Membranes were probed with anti-RTA antibodies and Western-blot analysis was performed. As shown in Figures 6(A) and 6(B), the differences between intracellular amounts of wild-type and P250A ricin increase with the time of incubation. Statistical analysis confirmed that the intracellular amount of ricin is dependent on the time and type of ricin (ANOVA II, time – F3,72=24, P<0.001; type of ricin – F1,72=37, P<0.001; for both factors F3,72=8, P<0.001). Moreover, differences between amounts of wild-type and P250A ricin after indicated times of toxin incubation are statistically significant (P<0.001); however, differences after 10 min are marginal.
It has been reported previously that proteolytic cleavage of the RTA and ricin holotoxin takes place mainly in early endosomes, additionally in late endosomes [6,7]. If degradation of the ricin P250A mutant is principally dependent on endosomal proteases, one could possibly observe differences in the amount of wild-type and P250A ricin recycled back to the cell surface. To analyse recycling, 125I-labelled wild-type and mutant P250A ricin were used. The amount of the holotoxin P250A mutant recycled back to the cell surface was significantly decreased in comparison with wild-type ricin (Figure 6C). These results support the hypothesis that ricin containing mutation in the RTA could be more extensively degraded in endosomes than wild-type ricin and therefore less P250A mutant is recycled back to the cell surface. This observation also supports the hypothesis that the lower amount of the sulfated holotoxin P250A mutant results from the decreased transport of RTAP250A:RTB to the Golgi apparatus, caused by a lower endosomal amount of P250A holotoxin accessible for endosome-to-Golgi transport.
Mutation in the hydrophobic region of RTA affects its retrotranslocation from the ER to the cytosol
To investigate whether retrotranslocation of RTA from the ER to the cytosol is affected by the mutation in the C-terminal region of the ricin A-chain, transport of the P250A mutant to the cytosol was studied. Cells incubated with wild-type or modified ricin for 3 h were subjected to permeabilization with digitonin to separate the cytosolic fraction from the ER membranes. In a control experiment, we tested that ER proteins remained in the ER after permeabilization. We studied the distribution of cnx, an ER membrane protein, and crt, a soluble ER protein. As shown in Figure 7(H), these protein markers were not released from the ER to the cytosolic fraction, indicating that ER membranes remained intact during digitonin treatment.
Figures 7(A) and 7(B) show an 8–9-fold reduction in the amount of modified RTA present in the cytosol fraction in comparison with wild-type RTA. Approx. 3-fold less efficient retrotranslocation of RTAP250A from the ER to the cytosol was observed in comparison with wild-type RTA (Figures 7C and 7D). Importantly, the difference in the amount of free wild-type RTA and RTAP250A in the ER is not higher than 3-fold (Figures 3A, 7C and 7E), similar to differences observed for the holotoxin present in the Golgi complex. Considering the fact that the reduction of wild-type and modified holotoxin in the ER is the same (Figures 3A and 3B), we conclude that the efficiency of the transport of modified RTA from the Golgi complex to the ER is the same as for wild-type ricin. Thus the mutation in the hydrophobic region of RTA affects its transport from the ER to the cytosol, and the observed decrease in cytosolic RTA P250A cannot be explained by reduced transport to the ER.
The difference in the cytoplasmic accumulation of wild-type and P250A RTA is not due to increased proteasomal degradation of RTAP250A. A proteasomal inhibitor, lactacystin, increases the amounts of both wild-type and modified ricin equally (Figures 7F and 7G). These results are consistent with the results of cytotoxicity experiments performed in the presence of lactacystin (Figure 2D).
The combination of a higher degradation level of the ricin P250A mutant in endosomes and the lower retrotranslocation of RTAP250A from the ER to the cytosol might explain the estimated 9-fold lower toxic effect of the holotoxin P250A mutant than of wild-type ricin (Figure 2).
Retrotranslocation of the RTA mutant from the ER to the cytosol is EDEM1-independent
It has been shown previously that the ER chaperone protein, EDEM1 plays a direct role in ricin transport from the ER to the cytosol . To further investigate the mechanism behind the reduced translocation of the P250A mutant to the cytosol, experiments with EDEM1-transfected cells and cells transfected with siRNA (small interfering RNA) vectors against EDEM1 were performed. Retrotranslocation of the wild-type and RTA P250A mutant was studied in cells permeabilized with digitonin, as described above. Two different vectors expressing shRNA (short hairpin RNA), against both mouse and human EDEM1, have been described previously . Figure 8(C) demonstrates a very efficient knockdown of mouse EDEM1 for both constructs. Reduction in the expression of endogenous human EDEM1 was confirmed by real-time RT–PCR (reverse transcription–PCR), which revealed a 70–95% reduction in the mRNA level of this gene (results not shown).
In cells with reduced amounts of EDEM1, the level of wild-type RTA present in the cytosol was decreased (Figures 8A and 8B), in agreement with previous findings . This confirms a direct role of EDEM1 in RTA transport to the cytosol. Importantly, the transport of the RTA P250A mutant from the ER to the cytosol seems to be EDEM1 independent, as retrotranslocation assays showed no significant differences in the cytosolic amount of RTAP250A in cells with reduced amounts of EDEM1 in comparison with cells transfected with an empty control vector (Figures 8A and 8B).
To determine whether the results from the retrotranslocation assays correspond to the inhibition of protein synthesis caused by the translocated toxin, control cells and cells transfected with siRNA vectors against EDEM1 were incubated with increasing concentrations of wild-type ricin or mutant P250A. We observed an ~2-fold protection (statistically significant, P< 0.001) of siRNA-transfected cells against wild-type ricin in comparison with control cells and no significant influence of EDEM1 down-regulation on the cytotoxicity of P250A ricin (Figures 8D and 8E). These results are in agreement with the suggestion that the RTA P250A mutant lacks EDEM1 contribution in the transport to the cytosol. This might explain the observed decrease in the retrotranslocation of RTAP250A from the ER to the cytosol in comparison with wild-type ricin (Figure 7).
Permeabilization assays in cells with overproduction of EDEM1 revealed a decrease in both wild-type RTA and mutant P250A present in the cytosol, when compared with cells transfected with control cDNA (Figures 9A and 9B). It was shown previously that high expression of EDEM1 increases ERAD  by promoting extraction of misfolded proteins from the cnx cycle [16,24]. Thus in EDEM1-transfected cells, the translocon is mainly occupied by misfolded proteins trans-ported to the cytosol for degradation and access of ricin to the translocon might be inhibited. This suggestion was confirmed in experiments with overproduction of model misfolded proteins . EDEM1 promotes RTA transport from the ER to the cytosol, but retrotranslocation assays showed a decreased amount of RTA present in the cytosol in EDEM1-transfected cells due to ERAD acceleration . Therefore, in cells with high levels of EDEM1, the availability of the translocon might be limited not only for wild-type but also for RTAP250A, despite the different influence of EDEM1 on RTA and RTAP250A retrotranslocation to the cytosol. In agreement with this idea, we observed protection against both wild-type ricin and mutant P250A in cells with overproduction of EDEM1 (Figures 9C and 9D).
Together, these results support the hypothesis that mutation in the hydrophobic region of RTA disrupts EDEM1-dependent transport of RTA to the cytosol.
P250A mutation impairs interaction between EDEM1 and RTA
It was shown previously that EDEM1 interacts directly with the ricin holotoxin . To study the interactions between RTA and EDEM1 and between RTAP250A and EDEM1, we cloned and purified the RTA mutant with a His-tag (see the Experimental section). As shown in Figure 10, EDEM1 interacts with the RTA. This new finding more precisely characterizes the previously observed interaction between EDEM1 and holotoxin . Interestingly, a significant decrease in RTAP250A binding to EDEM1 was observed when compared with wild-type RTA (Figure 10). These results might explain EDEM1-independent transport of RTAP250A to the cytosol. Importantly, CD analysis of RTA and RTAP250A with His-tags confirmed altered secondary structure of RTAP250A, presented above (Figure 1B).
Our pull-down assays were confirmed by in vivo experiments. We co-immunoprecipitated wild-type ricin or P250A ricin with anti-HA antibodies from lysates of EDEM1-HA-transfected cells or cells transfected with a control empty vector. As shown in Figure 10(C), both types of ricin co-immunoprecipitate with EDEM1. Importantly, much more wild-type ricin immunoprecipitates with EDEM1 in comparison with P250A ricin (Figures 10C and 10D). The level of interaction between wild-type ricin and EDEM1 and P250A ricin and EDEM1 differs in comparison with pull-down assays. This is due to the already decreased amount of P250A reaching the Golgi complex and consequently the ER (Figure 3).
These results additionally confirm that the interaction between P250A ricin and EDEM1 is decreased in comparison with wild-type ricin. Since the P250A mutation slightly changes the conformation of the ricin A-chain into a more helical structure, it is likely that the precise structure of an ERAD substrate might determine its recognition by EDEM1.
Increased knowledge about protein toxins from plants and bacteria is important both for the development of new therapeutic strategies and for characterization of basal mechanisms in cell biology. The results in the present study show that point mutation (P250A) in the hydrophobic region of the RTA influences endosomal–lysosomal degradation of the toxin, as well as RTA retrotranslocation from the ER to the cytosol. Our results indicate that the interactions between structurally changed RTAP250A and EDEM1 are impaired. This observation may significantly contribute to the general understanding of the recognition of protein substrates by EDEM1.
Our results show that mutated P250A ricin is 9-fold less toxic than wild-type ricin to Vero and HEK-293 cells. It has been demonstrated previously that the P250A mutant has significantly lower activity than wild-type ricin in Vero cells , but the reason behind the reduced cytotoxicity was not investigated. Differences in the activity of P250A and wild-type ricin reported previously  were higher than those described in the present study. Slightly different cell culture conditions, some variability in protein synthesis measurements (e.g. in the present study, results for single radioactivity measurement come from much higher amounts of cells) and different sublines of Vero cells may explain such a discrepancy.
Results from experiments performed in the presence of bafilomycin A, which prevents the acidification of endosomes and lysosomes, suggested an increased endosomal–lysosomal degradation of the mutated P250A holotoxin. Furthermore, the kinetics of degradation suggested that P250A ricin degradation starts already in the endosomes, as previously reported for wild-type ricin [6,7]. It has been demonstrated that the proteases cathepsin B and cathepsin D are responsible for RTA degradation in early endosomes in macrophages; cathepsin B is responsible for toxin cleavage at both neutral and acidic pH . Our results clearly indicate that cathepsin B and additionally cathepsin D are involved in increased P250A ricin degradation. Distinct endosomal proteases appear to be active in different cell types. In our study with HEK-293 cells, both wild-type ricin and mutated P250A degradation is at least partially dependent on low pH, since bafilomycin A treatment increased intracellular amounts of both toxins and decreased their degradation.
It was observed previously that addition of the RTB resulted in the protection of RTA from proteolytic activities of lysosomes and cathepsins . This protective role of RTB is partially responsible for the higher toxicity of ITs containing ricin holotoxin in comparison with ITs containing only RTA. Re-association of RTAP250A with RTB produces a holotoxin with a reducible disulfide bond; it produces a toxin that is equally bound to the cell surface and equally endocytosed as wild-type ricin. However, due to the structural changes in the mutated P250A RTA, RTAP250A:RTB can have altered conformation in comparison with the wild-type holotoxin. This altered conformation and/or localization of RTB in relation to RTAP250A might influence its degradation. It should be noted that in vitro degradation of RTA and RTAP250A by pronase at either pH 5.5 and 6.5 gives similar degradation patterns (results not shown).
Our results demonstrate that a point mutation (P250A) in the RTA decreases its retrotranslocation from the ER to the cytosol. A lack of EDEM1 contribution in this transport might be partially responsible for the observed effect. EDEM1 promotes RTA retrotranslocation to the cytosol ; however, this protein may not be the only one that facilitates RTA transport to the cytosol. In addition to the EDEM1 variant protein , EDEM2 and EDEM3 homologues also exist in the ER lumen [33–35]. Both EDEM2 and EDEM3 could be involved in ricin retrotranslocation. It is also possible that ricin can use not only Sec61p but also other channels/translocons for its transport to the cytosol. Thus one can expect that RTAP250A retrotranslocation from the ER to the cytosol might be influenced by additional factors; a lack of the assistance of EDEM1 may not be the only reason for the reduction of RTAP250A transport to the cytosol.
It has been demonstrated that RTA binds directly to the membrane surface; at the physiologically relevant temperature of 37 °C, the membrane-bound RTA loses some α-helical structures, undergoing the conformational change that exposes its C-terminal region to the membrane interior . Such insertion into the lipid bilayer might represent an early step in RTA translocation through the ER membrane. It is possible that the P250A mutant possessing an elevated level of α-helices is unable to undergo additional conformational changes allowing it to be stably inserted into the ER membrane. This might be another limiting step in RTAP250A retrotranslocation to the cytosol.
EDEM1 substrate recognition and sorting to the ERAD pathway are still poorly defined. It has been proposed that EDEM1 acts either as mannosidase that produces de-mannosylated glycoproteins or as the receptor that recognizes, binds and directs mannose-trimmed proteins for ERAD by extracting them from the cnx cycle [16–19]. On the other hand, it has been proposed that EDEM1 substrate recognition might be glycan independent. EDEM1 directly interacts with non-glycosylated ricin, this interaction is independent of glycans presented on RTB . Hebert and co-workers  demonstrated that EDEM1 binding does not require the trimming of substrate glycans or even ERAD substrate glycosylation, thus suggesting that EDEM1 probably recognizes misfolded regions of aberrant proteins. Similarly, EDEM1 binding to mutant P23H rod opsin was independent of mannose trimming . In the present study, we demonstrate that point mutation in the hydrophobic region of the RTA decreases the interaction between RTA and EDEM1. It has been suggested that EDEM1 interacts with proteins via hydrophobic domains . Hydrophobicity of the mutated P250A C-terminal domain of RTA is not changed dramatically in comparison with wild-type RTA. However, a point mutation in this region changes the secondary structure of RTA into more a helical one. This suggests that protein structure might influence EDEM1–protein substrate recognition. This hypothesis has to be confirmed in further experiments. Understanding the mechanisms of recognition and degradation of misfolded proteins synthesized in the ER is one of the fundamental issues in cell biology.
The potency of the protein toxin ricin can be combined with the specificity of various targeting moieties to yield ITs used in cancer therapy. However, a major disadvantage is the dose-limiting toxicity associated with ricin-conjugated ITs, which leads to VLS (vascular leak syndrome). It is characterized by hypoalbuminaemia, peripheral oedema and pulmonary oedema in most severe cases . VLS is caused by RTA, which damages vascular endothelial cells by increasing their permeability in a time- and dose-dependent manner . The exact mechanism by which RTA mediates VLS in vivo remains to be determined, but there is evidence suggesting that direct cytotoxicity of RTA leads to VLS in vivo . Thus, as an alternative, the use of toxins that are less cytotoxic to endothelial cells are proposed. Ricin with a mutated hydrophobic region of RTA seems to be a good candidate for such studies. It is still potent enough to be used in ITs, on the other hand it is possible that VLS caused by this toxin would be strongly limited. The influence of RTAP250A on endothelial cell permeability should be determined.
In conclusion, the results presented in the present study contribute to our general understanding of ricin intracellular transport and mechanism of the recognition of misfolded proteins in the ER. This knowledge might bring new solutions into therapies of severe human diseases.
Monika Słomińska-Wojewódzka and Kirsten Sandvig conceived the project and designed the experiments. Monika Słomińska-Wojewódzka and Iwona Sokołowska performed all experiments and collected the data. Sébastien Wälchli cloned RTA-His and established the purification protocol. Sébastien Wälchli and Monika Słomińska-Wojewódzka performed the confocal microscopy experiments. Monika Słomińska-Wojewódzka, Kirsten Sandvig and Iwona Sokołowska analysed the data. Monika Słomińska-Wojewódzka and Kirsten Sandvig wrote the manuscript together with Sébastien Wälchli, Grzegorz Węgrzyn and Iwona Sokołowska. All authors read and approved the final version of the manuscript.
This work was supported by the Ministry of Science and Higher Education [grant number 3682/P01/2006/32], Foundation for Polish Science [grant number ‘Homing’ HOM/12/2007] and by the University of Gdańsk [grant number 1480-5-0344-6].
We are grateful to Professor Kazuhiro Nagata and Dr Nobuko Hosokawa (Institute for Frontier Medical Sciences, Kyoto University, Kyoto, Japan) for cDNA encoding the mouse EDEM1. We would like to thank Dr Danuta Augustin-Nowacka (Faculty of Chemistry, University of Gdańsk, Gdańsk, Poland) for CD measurements and analysis. We are grateful to Dr Sigrid S. Skånland (Institute of Biochemistry II, Goethe University School of Medicine, University Hospital, Frankfurt, Germany) and to Dr Anna Herman-Antosiewicz (Department of Molecular Biology, University of Gdańsk, Gdańsk, Poland) for help in the confocal microscopy images preparation. We thank Dr Anna Kawiak (Department of Biotechnology, Intercollegiate Faculty of Biotechnology, University of Gdańsk, Gdańsk, Poland) for critical reading of the manuscript.
Abbreviations: cnx, calnexin; crt, calreticulin; DTT, dithiothreitiol; ER, endoplasmic reticulum; EDEM1, ER degradation-enhancing α-mannosidase I-like protein 1; ERAD, ER-associated protein degradation; FBS, fetal bovine serum; HA, haemagglutinin; HBS, Hepes-buffered saline; HEK-293, cells, human embryonic kidney cells; HRP, horseradish peroxidase; IT, immunotoxin; Ni-NTA, Ni2+-nitrilotriacetate; RTA, ricin A-chain; RTB, ricin B-chain; siRNA, small interfering RNA; TGN, trans-Golgi network; TGN46, trans-Golgi network protein, 46 kDa; VLS, vascular leak syndrome
- © The Authors Journal compilation © 2011 Biochemical Society