All chemicals were purchased from Sigma-Aldrich (St. Louis, MO, USA) if not otherwise noted. H₂³⁵SO₄ and [³H]Leucine were purchased from Hartman Analytic (Braunschweig, Germany), D-[2-³H(N)]-mannose and Na¹²⁵I were from Perkin Elmer (Boston, MA, USA). ONO-RS-082 (ONO) was purchased from ENZO Life Sciences (Plymouth, PA, USA), and bromoenol lactone (BEL) from Sigma-Aldrich. Rabbit anti-ricin antibody was from Sigma Aldrich, sheep anti-TGN46 antibody was from AbD Serotec (Oxford, UK), and mouse anti-GM130 was from BD Bioscience (San Jose, CA, USA). Ricin-sulf-1 (RS1), ricin with a modified ricin A-subunit containing a tyrosine sulfation site, and ricin-sulf-2 (RS2), ricin with a modified ricin A-subunit containing a tyrosine sulfation site and three partially overlapping N-glycosylation sites in the C-terminal, was produced and purified as described earlier [33]. Plasmids encoding TPST-1-EGFP and TPST-2-EGFP were a kind gift from Dr. Ludger Johannes (Institut Curie, Paris, France).

HEp-2 cells were maintained in Dulbecco’s modified Eagle’s medium (Invitrogen, Carlsbad, CA, USA) supplemented with 10% fetal calf serum, 100 units/mL penicillin and 100 µg/mL streptomycin, at 37 °C under 5% CO₂. For cytotoxicity and endocytosis assays, the cells were seeded out in 24-well plates at a density of 5 × 10⁴ cells/well one day before the experiment. For the sulfation and mannosylation assays, cells were seeded in 6-well plates at a density of 2 × 10⁵ cells/well, and for confocal microscopy studies the cells were seeded in 6-well plates with two coverslips per well at a density of 1 × 10⁵ cells/well, one day before the experiment. For transfection with pTPST-1-EGFP and pTPST-2-EGFP, cells on coverslips were transfected using Fugene 6 according to the manufacturer’s instructions (Roche Diagnostics, Mannheim, Germany) one day before use in confocal microscopy studies.

Cells grown in 24-well plates were washed twice in leucine-free HEPES-buffered medium before addition of inhibitors at indicated concentrations for 30 min. Increasing concentrations of ricin (1–1000 ng/mL) were added, and the cells incubated for 3 h at 37 °C. The medium was then replaced with leucine-free HEPES-buffered medium containing 2 µCi/mL [³H]Leucine, and the cells incubated further for 20 min. The proteins were precipitated with 5% (w/v) trichloracetic acid (TCA), washed once in 5% (w/v) TCA, and then dissolved in 0.1 M KOH. The incorporation of radioactively labeled leucine was quantified, and IC50 calculated as the concentration of toxin giving a 50% reduction in protein synthesis. Fold protection was then calculated as the ratio between IC50 for inhibitor-treated cells compared to control cells.

The cells were washed twice with sulfate-free medium supplemented with 2 mM L-glutamine, followed by incubation with 0.2 mCi/mL H₂³⁵SO₄ in sulfate-free medium for 2.5 h at 37 °C. The cells were then incubated with inhibitors at the indicated concentrations for 30 min before addition of ricin-sulf-1 (6 µg/mL) for an additional 2 h. The medium was removed and the cells washed 2 × 5 min with 0.1 M lactose in HEPES-buffered medium, and once in cold PBS on ice before addition of 400 µL lysis buffer [0.1 M NaCl, 10 mM Na2HPO4, 1 mM EDTA, 1% Triton X-100, 60 mM octyl glycopyranoside] supplemented with Complete protease inhibitors (Roche Diagnostics, Mannheim, Germany). The lysate was cleared by centrifugation (8000 rpm, 10 min) and ricin-sulf-1 was immunoprecipitated with anti-ricin antibody, pre-bound to protein A-sepharose beads (Amersham Biosciences, Buckinghamshire, UK), overnight at 4 °C. The precipitate was separated by SDS-PAGE under reducing conditions, and blotted onto a PVDF membrane (Immobilon-P, Millipore, Billerica, MA, USA). The bands were detected by autoradiography and imaged and quantified using PharosFX scanner and Quantity One^® 1-D Analysis Software (BioRad Laboratories Inc, Hercules, CA, USA). The total amount of sulfated proteins was determined by TCA-precipitation of all ³⁵S-labeled proteins in the lysates.

The cells were washed twice with glucose-free DMEM, followed by incubation in glucose-free DMEM containing 1.8 µL/mL D-(+)-glucose solution (10%) and 0.1 mCi/mL D-[2-³H(N)]-mannose for 2.5 h at 37 °C. ONO was then added at indicated concentrations, in addition to 1 µg/mL swainsonine, and the cells incubated for another 30 min. Ricin-sulf-2 (6 µg/mL) was then added and incubation continued for another 2 h. Subsequently, the cells were treated following the sulfation assay protocol described above. The bands were detected using Kodak BioMax MS-film (Sigma-Aldrich), scanned and quantified by Quantity One^® 1-D Analysis Software (BioRad Laboratories Inc, Hercules, CA, USA). The total amount of mannosylated proteins was determined by TCA-precipitation of all ³H-labeled proteins in the lysates.

Cells were pre-incubated with 2.5 or 5 µM ONO for 30 min, before addition of ¹²⁵I-ricin for 2 h at 37 °C. Binding of ricin was measured as the amount of cell-associated ricin after washing with ice-cold buffer (0.14 M NaCl, 2 mM CaCl₂, 20 mM HEPES; pH 7.0). Endocytosed ricin was measured as the amount of ¹²⁵I-ricin that was not removed after 2 × 5 min washing with 0.1 M lactose in HEPES-buffered medium.

The cells grown on glass coverslips were washed once in HEPES-buffered medium before addition of 5 µM ONO for 30 min. Ricin-sulf-1 (~1 µg/mL) was added for 15 min, followed by 30 min chase. The cells were then washed twice with 0.1 M lactose in HEPES-buffered medium, and twice in PBS. The cells were fixed with 10% formalin-solution, and permeabilized in 0.1% Triton X-100. Blocking solution, 5% fetal calf serum in PBS, was then added for 30 min, before incubation with the appropriate primary and secondary antibodies. Fluorophore-labeled secondary antibodies were obtained from Jackson ImmunoResearch Laboratories (West Grove, PA, USA). The cells were mounted in Prolong^®Gold containing DAPI for staining of the nuclei (Molecular Probes, Eugene, OR, USA), and imaged on laser scanning confocal microscope LSM 510 DUO (Carl Zeiss, Jena, Germany). Images were prepared and analyzed with the LSM Image Browser software (Carl Zeiss) or ImageJ software (NIH, Bethesda, USA). Colocalization was quantified using the LSM Image Examiner (Carl Zeiss), and presented as percentage colocalization compared to control.

In order to examine the potential involvement of PLA₂ enzymes in the intracellular transport of ricin, we took advantage of the commonly used PLA₂ inhibitors ONO-RS-082 (ONO) and bromoenol lactone (BEL). First we investigated if treatment with ONO had any effect on the overall cytotoxicity of ricin in HEp-2. The cells were pretreated with two different concentrations of ONO, before exposure to increasing concentrations of ricin. Toxin response curves were created, and as can be seen in Figure 1A, pre-treatment with 5 µM ONO had a significant protective effect against ricin. As a quantitative measure of this effect, the fold protection was calculated as the concentration of ricin giving a 50% reduction in protein synthesis in treated cells compared to untreated cells (Figure 1C), and 5 µM ONO resulted in an almost 9-fold protection against ricin. Cytoplasmic PLA₂s can be divided into Ca²⁺-dependent and Ca²⁺-independent (iPLA₂) enzymes [20], and ONO is an inhibitor of both of these classes. BEL is an irreversible inhibitor of iPLA₂, and was also tested in ricin toxicity assays. At a concentration of 2.5 µM, which is commonly used to inhibit iPLA₂ activity [26,29], BEL did not have any protective effect against ricin (Figure 1B,C). These data then suggest that it is the Ca²⁺-dependent PLA₂ enzymes that are involved in the ricin intoxication process.

Shiga toxin (Stx) is a bacterial toxin which is also endocytosed and transported retrogradely from the Golgi to the ER, and finally the cytosol (see e.g., [34]). Interestingly, treatment with 2.5 or 5 µM ONO had no significant effect on the toxicity of Stx in these cells (Supplementary Figure S1A), indicating that Stx transport to the cytosol of HEp-2 cells is not dependent on PLA₂ activity.

Given the observed protection against ricin by the PLA₂ inhibitor ONO, we were interested in determining which step in the retrograde transport pathway that was impaired in the presence of this compound. As mentioned in the introduction, PLA₂ activity is important for endosomal membrane tubulation, and PLA₂ enzymes are involved in several endosomal trafficking events [26,28]. It could therefore be speculated that ricin transport from endosomes to the Golgi could be impaired in response to PLA₂ inhibition. In order to address this, we took advantage of a modified ricin molecule, ricin-sulf-1, with a tyrosine sulfation site attached C-terminally to the ricin A-chain. As sulfation is a modification that occurs mainly in the TGN [35], it is possible to measure the amount of internalized ricin reaching the Golgi apparatus using radioactively labeled sulfate. Interestingly, ricin transport to the Golgi is not impaired in the presence of inhibitors ONO (Figure 2A) or BEL (Figure 2B). Surprisingly, the amount of sulfated ricin was strongly increased in response to inhibitor treatment. Control experiments show that there was no effect on total protein sulfation in the cells. Quantification of the data shows that there is as much as an 8-fold increase in the amount of sulfated ricin-sulf-1 in response to 2.5 µM ONO and a 10-fold increase for 5 µM ONO. For BEL, the effect on ricin sulfation was weaker, but still giving a 2.5-fold increase in response to 10 µM BEL. These data indicate that ricin is being transported to the Golgi apparatus in presence of inhibitors, and they show that the degree of sulfation is actually increased after inhibitor treatment. The sulfation of Stx was not altered in response to ONO (Supplementary Figure S1B).

Increased ricin sulfation could be explained by an increase in the internalization of the toxin. In order to examine this possibility, endocytosis of ¹²⁵I-labelled ricin was analyzed. As can be seen from Figure 3A, there was no significant change in the amount of endocytosed ricin in response to ONO, demonstrating that PLA₂ activity is not required for ricin internalization.

As described in the introduction, ricin is endocytosed and retrogradely transported from early endosomes, through the Golgi apparatus and to the endoplasmic reticulum (ER), before it is translocated to the cytosol where it exerts its cytotoxic effect. The decreased toxicity of ricin in response to PLA₂ inhibition, in combination with the increased sulfation signal, led us to investigate if the transport of the toxin from the Golgi to the ER might be inhibited. To this end we used a modified ricin molecule, ricin-sulf-2, containing a modified A-chain with N-glycosylation sites in combination with radioactively labeled mannose. As mannosylation is an ER-specific modification, this gives a measure of how much of the internalized ricin is being transported to the ER. The data presented in Figure 3B shows a dramatic decrease in mannosylation of ricin-sulf-2 after treatment with 5 µM ONO, indicating that the transport of ricin to the ER is strongly impaired. Quantification shows that there is a close to 90% reduction in ricin mannosylation in the presence of 5 µM ONO. Control experiments show a moderate reduction (~30%) in total mannosylation in response to ONO.

The sulfation and mannosylation data suggest that PLA₂ activity is essential for Golgi-to-ER transport of ricin, and that PLA₂ inhibition thereby protects the cells against ricin intoxication. ONO and BEL are known to cause disruption of the Golgi cisternae [23,24], but as ricin is strongly sulfated in the presence of the inhibitors it seems evident that ricin is transported to the Golgi also under these conditions. To get a clearer picture of what is happening to the Golgi in response to PLA₂ inhibition, and where ricin is localized under these conditions, we performed confocal microscopy studies were we stained cells for ricin in combination with Golgi markers. In Figure 4A we show the localization of the trans-Golgi marker TGN46 and the cis-Golgi marker GM130 in untreated cells and cells treated with 5 µM ONO. As can be clearly seen from the images, both TGN46 and GM130 staining show that the inhibitor induces fragmentation of the Golgi as reported previously [23,24]. The quantification (Figure 4B) shows that the apparent colocalization between TGN46 and GM130 is decreased in response to ONO, indicating that the inhibitor induces a separation between the TGN and the cis-Golgi.

To analyze the localization of ricin under the same conditions, we stained cells for ricin in addition to TGN46 and GM130 (Figure 4C–E). As can be seen from the images (Figure 4C,D), and more clearly from the quantification of colocalization (Figure 4E), ricin is colocalized with TGN46 to the same extent both in the absence and presence of ONO, while there is a decrease in colocalization with GM130 upon treatment with the inhibitor (from approximately 30% before treatment to 20% after treatment). This indicates that ricin is able to be transported to the TGN, but cannot to the same extent be transported to the cis-Golgi. The localization of Stx was not altered in response to ONO (Supplementary Figure S2), demonstrating that the change in localization of ricin is not just a result of the ONO-induced change in Golgi structure. An accumulation of ricin in TGN over time could also be observed by confocal analysis (Supplementary Figure S3), supporting the idea that ricin is inhibited from being transported further along the retrograde transport route.

The increased sulfation seen in response to PLA₂ inhibition may be due to either more ricin being transported to the TGN, more efficient sulfation in the TGN, or an accumulation of ricin in the TGN as it is inhibited from being transported further on along the retrograde pathway and finally degraded. As the amount of total protein sulfation is not affected by the inhibitors, a general increase in the activity of sulfotransferases does not seem to be a likely explanation. To explore if the treatment with PLA₂ inhibitor is affecting the localization of the two main sulfotransferases TPST (Tyrosyl Protein SulfoTransferase)-1 and TPST-2, we expressed GFP-tagged versions of these enzymes in HEp-2 cells. The cells were transfected prior to treatment with ONO and exposure to ricin, the localization of the TPSTs was then examined by confocal microscopy, and the colocalization with Golgi-markers and ricin was quantified. As the images in Figure 5A (TPST-1) and 5B (TPST-2) show, the sulfotransferases are mainly localized to the TGN in untreated cells, as expected. Upon addition of ONO, their localization became more dispersed, as observed also for TGN46. The quantification (Figure 5C) shows that the colocalization of TPST-1 and TPST-2 with TGN46 is practically unchanged upon treatment. In addition, colocalization between the sulfotransferases and ricin was also unaffected. These data show that although the Golgi is dispersed upon treatment with ONO, both ricin and sulfotransferases are still localized to the TGN. These data then suggest that the observed increase in ricin sulfation is probably due to either increased transport to the TGN after endocytosis, or accumulation of the toxin in the TGN, or a combination of these.