The Proteasome Is a Molecular Target of Environmental Toxic Organotins

Background Because of the vital importance of the proteasome pathway, chemicals affecting proteasome activity could disrupt essential cellular processes. Although the toxicity of organotins to both invertebrates and vertebrates is well known, the essential cellular target of organotins has not been well identified. We hypothesize that the proteasome is a molecular target of environmental toxic organotins. Objectives Our goal was to test the above hypothesis by investigating whether organotins could inhibit the activity of purified and cellular proteasomes and, if so, the involved molecular mechanisms and downstream events. Results We found that some toxic organotins [e.g., triphenyltin (TPT)] can potently and preferentially inhibit the chymotrypsin-like activity of purified 20S proteasomes and human breast cancer cellular 26S proteasomes. Direct binding of tin atoms to cellular proteasomes is responsible for the observed irreversible inhibition. Inhibition of cellular proteasomes by TPT in several human cell lines results in the accumulation of ubiquitinated proteins and natural proteasome target proteins, accompanied by induction of cell death. Conclusions The proteasome is one of the molecular targets of environmental toxic organotins in human cells, and proteasome inhibition by organotins contributes to their cellular toxicity.


Research
In eukaryotes, more than 80% of intra cellular proteins are degraded through the ubiquitin/proteasomedependent pathway (Ciechanover 1998;Dou and Li 1999;Goldberg 1995). The ubiquitin/ proteasome dependent pathway plays an essential role in antigen presentation, cellular aging, apoptosis, and other major cellular pro cesses. The cellular proteasome, commonly called 26S proteasome, is composed of two 19S regu la tory particles and a 20S core par ticle. The latter is a multicatalytic threonine protease with at least three distinct catalytic activities: chymotrypsin (CT)like (cleavage after hydrophobic residues mediated by the β5 subunit), trypsinlike (cleavage after basic residues by the β2 subunit), and peptidyl glutamyl peptidehydrolyzing (PGPH)like (cleavage after acidic residues by the β1 sub unit) (Goldberg 1995;Groll et al. 1997). Inhibition of proteasome CTlike activity by various compounds is associated with cell apoptosis Lopes et al. 1997).
Since the 1960s, organotins, especially tri phenyltin (TPT) and tributyltin (TBT), have been extensively used as antifouling boat paints, polyvinyl chloride stabilizers, agricultural pesti cides, and industrial catalysts. Consequently, organotin contamination is found in various environmental media (Fent 1996). Because of their lipophilic property, organotins can be accumulated through the food chain and reach higher concentrations in top predators. For example, levels of organotins in stranded whales reached 1.0-1.1 mg/kg (Harino et al. 2007) and in the liver of harbor porpoises reached 68-4,605 mg/kg (Strand et al. 2005). Organotin chemicals have also been found in the tissues of humans contaminated by organo tin insecticides or food (Kannan et al. 1999).
Exposure to TPT or TBT can also affect sex differentiation, resulting in masculinization of females or infertility in males (McAllister and Kime 2003;Smith 1996). TBT at low concen trations is toxic to cortical neurons by triggering glutamate excitotoxicity (Nakatsu et al. 2006). TBT also induces the differentiation of adipo cytes in vitro and increases adipose mass in vivo, perhaps through activation of the retinoid X receptor and the peroxisomeproliferatoractivated receptor γ (Grun et al. 2006). Both TPT and TBT can interfere with the cytotoxic function of natural killer (NK) cells (Snoeij et al. 1987), associated with increased cancer incidence. However, the toxicologic mechanism for organotin compounds is not completely understood, and the essential cellular target of organotins has not been identified.
The authentic proteasome inhibitor clasto lactacystin βlactone contains an electrophilic ester bond carbon that is responsible for its bio logical inhibition of the proteasome (Fenteany et al. 1995). We hypothesize that the electro philic tin (Sn) atom in organotins could also be attacked by the O γ atom of the Nterminal threo nine (Thr1) of the proteasome β5 subunit, causing irreversible inhibition. This hypothe sis is supported by the present results from in silico docking, by in vitro proteasome activ ity assay using purified 20S proteasomes and human breast cancer MDAMB231 cells and human peripheral blood Jurkat T cells treated with TPT and other organotins, by analysis of CTlike activity, β5 proteasome subunit expression, and Sn levels in isolated proteasome complexes from the treated cells.
Background: Because of the vital importance of the proteasome pathway, chemicals affecting proteasome activity could disrupt essential cellular processes. Although the toxicity of organotins to both invertebrates and vertebrates is well known, the essential cellular target of organotins has not been well identified. We hypothesize that the proteasome is a molecular target of environmental toxic organotins. oBjectives: Our goal was to test the above hypothesis by investigating whether organotins could inhibit the activity of purified and cellular proteasomes and, if so, the involved molecular mechanisms and downstream events. results: We found that some toxic organotins [e.g., triphenyltin (TPT)] can potently and preferentially inhibit the chymotrypsin-like activity of purified 20S proteasomes and human breast cancer cellular 26S proteasomes. Direct binding of tin atoms to cellular proteasomes is responsible for the observed irreversible inhibition. Inhibition of cellular proteasomes by TPT in several human cell lines results in the accumulation of ubiquitinated proteins and natural proteasome target proteins, accompanied by induction of cell death. conclusions: The proteasome is one of the molecular targets of environmental toxic organotins in human cells, and proteasome inhibition by organotins contributes to their cellular toxicity. key words: cell death, molecular target, organotins, proteasome, proteasome inhibitors, TPT.  MO, USA).
Cell cultures, cell extract preparation, and Western blot analysis. We cultured human breast cancer MDAMB231 cells and human peripheral blood Jurkat T cells and prepared wholecell extracts as previously described . Western blot analysis using the enhanced chemiluminescence reagent was per formed as previously described .
Proteasome activity assays using purified 20S proteasomes in intact cells. We measured the inhibition of purified 20S proteasomal activity ) and 26S proteasomal activity in living intact cells (Chen et al. 2005) as described previously.
Proteasome and caspase-3 activity assays using cell extracts. The prepared wholecell extracts (10 µg per sample) from treated cells were incubated with the appropriate fluorogenic peptide substrates in 100 µL assay buffer at 37°C for 2 hr. We measured the release of the 7amino4methylcoumarin (AMC) groups as previously described .
Trypan blue dye exclusion assay and terminal deoxynucleotidyl transferase dUTP nick end-labeling (TUNEL) assay. We used the trypan blue dye exclusion assay to ascertain cell death in Jurkat T cells treated with TPT or TBT as described previously (Chen et al. 2005). The TUNEL assay using TPTtreated cells was performed with an APODirect kit to determine the extent of DNA strand breaks measured by flow cytometry according to the manufacturer's instructions Chen et al. 2005).
Assay for interaction of TPT and purified 20S proteasome. We incubated purified rab bit 20S proteasomes (4 µg per reaction) with 10 µmol/L TPT or the control solvent metha nol in 400 µL TrisHCl buffer (25 mmol/L, pH 7.5) overnight at 4°C. The reaction mix ture was then transferred to the insert cup of UltrafreeMC centrifugal filter unit (MW cut off, 5 kDa), which had been prewashed with 400 µL 1% bovine serum albumin,1% sucrose solution to minimize proteasome binding to the unit, and the mixture centrifuged at 7,000 rpm for 1.5 hr at 4°C, which resulted in about 80 µL of the proteasome preparation in the insert cup. Then 320 µL TrisHCl buffer was added to the insert cup containing the proteasome prepa ration, followed by centrifugation again. This wash-filtration procedure was repeated four times to remove free TPT. The proteasome preparation in the insert cup was diluted to 100 µL with the TrisHCl buffer and transferred to a storage tube. The insert cup was then washed with 50 µL TrisHCl buffer and combined with the proteasome fraction. We used the prepared proteasome fraction for the proteasome activity and Western blot assays.
Assay for interaction of TPT and cellular 26S proteasome. We treated MDAMB231 cells with 10 µmol/L TPT or control solvent methanol for 2 hr, followed by preparation of wholecell extracts and the 26S proteasome immunoprecipitates. In brief, 100 µL agarose beads immobilized with mouse monoclonal antibody to the 20S proteasome subunit α2, or agarose beads immobilized with aprotinin (as a nonspecific binding control), were equili brated with 1 mL buffer A (TrisHCl buffer, 25 mmol/L, pH 7.5) and then incubated with 100 µL of the prepared cell extracts (~ 800 µg protein per preparation) overnight at 4°C in 700 µL buffer A. The reaction mixtures were centrifuged at 3,000 rpm for 1 min, the supernatants were removed, and the affinity matrixes were washed three times with 1 mL buffer A. The mixtures were resuspended with 400 µL buffer A and separated into two ali quots (200 µL each). One aliquot was diluted to 400 µL and incubated with SucLLVY AMC (40 µmol/L) fluorogenic substrate at room temperature for 4 hr. After centrifugation at 3,000 rpm for 1 min, the supernatant was transferred to a 96well plate for measurement of proteasome activity; the affinity matrix was resuspended with 100 µL 2× sodium dodecyl sulfate sample buffer, followed by Western blot analysis using proteasome β5 antibody. Another aliquot (200 µL) of the prepared proteasome immunoprecipitates was digested with an equal volume of 6 mol/L HCl for determination of the total Sn by inductively coupled plasmamass spectroscopy (ICPMS).

ICP-MS analysis of Sn bound to proteasome.
We determined total Sn levels in the prepared proteasome complexes with an Agilent 7500ce ICPMS (Agilent Technologies, Santa Clara, CA, USA). The operating conditions were as follows: radiofrequency power, 1,240 W; Argon plasma gas (15 L/min), carrier gas (0.8 L/min), and makeup gas (0.27 L/min); sampling depth, 8.5 mmol/L; operation mode, shield torch; acquired mass, 118; points/mass, 3. In silico models for TPT binding to proteasome β5 subunit. Because 99% of TPT is present as the neutral hydroxylcomplex form (TPTOH) in physiologic conditions (pH 7.5) (White and Tobin 2004), the chemical struc ture of TPT was selected as (C6H5)3SnOH for docking studies. The crystal structure of the proteasome β5 subunit, the docking param eters, docking methods, and output analysis methods were the same as described previously (Smith et al. 2004). In brief, we first refined the molecule by performing an optimized geometry calculation saved in Protein Data Bank (PDB) files using the conversion filters in CAChe software (v 6.1) (Fujitsu America Inc., Beaverton, OR, USA). The output PDB files were imported into AutoDock 3.0 soft ware (The Scripps Research Institute 2008) for the in silico binding analysis to the proteasome β5 subunit. We chose the crystal structure of the proteasome β5 subunit of the eukary otic yeast 20S proteasome (PDB reference no. 1JD2; RCSB Protein Data Bank 2008), which is similar to human 20S proteasome (Smith et al. 2004). We defined the Sn atom of TPT as a rigid atom and limited the dock ing space to a 20 × 20 × 20 Å box centered on the β5catalytic Nterminal threonine that was prepared as an energyscoring grid. The output from AutoDock was used for docking model studies with PyMOL software (Smith et al. 2004).

Inhibition of purified 20S proteasomes by organotins.
We hypothesized that the pro teasome is a major target of organotins. To test this hypothesis, we first investigated the inhibition potency of phenyltins and butyl tins to purified 20S rabbit proteasomal activ ity. Among all the phenyltins tested (MPT, DPT, TPT, and TePT), TPT was the most potent inhibitor against proteasomal CTlike activity, with a halfmaximal inhibitory con centration (IC 50 ) of 3.5 µmol/L ( Figure 1A). At 25 µmol/L, TPT and DPT caused 95% and 29% inhibition, respectively, whereas MPT and TePT had 15% or less inhibition ( Figure 1A). When we tested butyltins with the purified 20S proteasome, we found a simi lar order of the CTlikeinhibitory activity: TBT (IC 50 = 15.6 µmol/L) > DBT > MBT, TeBT ( Figure 1B).
To investigate whether TPT specifically inhibits proteasomal CTlike activity, we examined its effects on the PGPHlike and trypsinlike activities of purified 20S protea somes. At 10 µmol/L, TPT inhibited CTlike and PGPHlike activity of the purified 20S proteasomes by 79% and 44%, respectively, but had no inhibitory effect against trypsin like activity ( Figure 1C). Similarly, TBT had much less inhibitory effect on PGPHlike and trypsinlike activities of purified 20S proteasomes (data not shown). Therefore, it appears that organotins such as TPT and TBT preferentially inhibit proteasomal CTlike activity over other activities.
To further investigate the nature of TPT mediated proteasome inhibition, we incu bated the purified 20S proteasomes overnight at 4°C with either 10 µmol/L of TPT or control solvent. CTlike activity assays using an aliquot of each preparation indicated that TPT inhibited proteasome activity by 57% under these experimental conditions ( Figure 1D, "before spin"). Most of the reac tion mixtures were then washed several times with an UltrafreeMC centrifugal filter unit (MW cutoff, 5 kDa) to remove small mole cules, including free TPT. The proteasomal activity assay shows that repeated washing and filtration of the TPTproteasome mixture did not change the outcome: TPT still caused 51% inhibition of proteasomal CTlike activity ( Figure 1D, "after spin"). The lower level of the proteasomal activity detected in the TPT-20S proteasome mixture was not due to decreased levels of 20S proteasomes because an equal level of β5 subunit protein was detected by Western blotting in the two preparations ( Figure 1E). This result indicates that TPT is either a tightbinding or an irre versible inhibitor of the proteasome.
Computational study of TPT-proteasome β5 interaction. In order to build a model to explain how TPT inhibits proteasomal CTlike activity, we performed automated docking studies. The docking results indi cate that TPT has a major docking mode that repeated for 47 of 100 runs (47% probabil ity). The distance from the Sn atom to the O γ on Thr1 was 2.97 Å, and one of the phenyl rings of TPT was located within the S 1 hydro phobic pocket of β5 subunit ( Figure 2A). It is well documented that the Thr1 O γ atom could be activated by Thr1 N directly or via a neighboring water molecule, and reacts with  (Groll et al. 1997). Because the Sn atom in TPT was close to the Thr1 O γ atom, we hypothe sized that the Thr1 O γ atom might perform a nucleophilic attack on the Sn atom and form a coordinate bond ( Figure 2B). It has been reported that with the attack of a nucleophilic ligand, the Cl ligand on the Sn atom of TBT or TPT could be replaced by the nucleophilic ligand, or the hybridization state of the Sn atom could change from sp3 to dzsp3, which allows the Sn atom to form a new coordinate bond with the nucleophilic ligand (Burda et al. 2002). After the nucleophilic attack by Thr1 O γ , the TPT-Thr1 complex could be present in two possible conformations: tetra hedral (Sn in sp3 hybridization) and trigonal bipyrami dal (Sn in dzsp3 hybridization). To deter mine whether the hybrid orbit orientation of Sn atom and the threedimension space around Thr1 O γ could allow the formation of any of these conformations, we performed further docking studies. To mimic the coor dinate complex of TPT-Thr1, a threonine amide group was used as one of the ligands for Sn. The complex structure was refined by performing an optimized geometry cal culation in MOPAC using PM5 parameters in the CAChe software. After optimization of the geometry, the threonine amide ligand was removed and the remaining structure was docked into the proteasome β5 subunit with AutoDock 3.0 software. The docked results indicate that TPT has one major trigo nal bipyramidal conformation (Sn in a five coordinate form; Figure 2C) that allows the formation of a coordinate bond. In contrast, the tetrahedral conformation (Sn in a four coordinate) does not give an orientation that allows the Sn and Thr1 O γ to form a coordi nate bond (data not shown).
In the major trigonal bipyramidal confor mation mode that repeated in 63 of 100 runs (63% probability), the OH ligand and two phenyl rings of TPT were on the equatorial orbit, whereas the third phenyl ring of TPT and the Thr1 O γ of β5 was on the axis orbit ( Figure 2C). The distance between Sn and O γ was 2.81 Å, and the line between Sn and O γ was basically on the axis orbit orientation ( Figure 2C), supporting that a coordinate bond could form between Sn and O γ . Also, one of the TPT hydrophobic phenyl ligands was oriented in the S 1 pocket of β5 subunit ( Figure 2D). The hydrophobic portion of the aromatic phenyl ring is oriented in the mid dle of the S 1 pocket, with distances of 4.21 and 3.48 Å, respectively, to the side chains of Ala49 and Lys33 ( Figure 2D). In addition, the sidewalls of the S 1 pocket, which pos sibly interact with TPT hydrophobically, are created by Met45, Ala20, and Val31 with distances of 3.26, 3.43, and 3.52 Å, respec tively ( Figure 2D). We also found that the distance between the amide hydrogen of Thr 21 and the oxygen on the hydroxide ligand of TPT is only 2.00 Å ( Figure 2C), indicating the possible formation of a hydrogen bond. Therefore, the irreversible inhibition nature of TPT is supported by the docking results, the possible formation of a coordinate bond between Sn of TPT and Thr1 O γ , blockage of the S 1 pocket by a TPT phenyl ring, the possible hydrophobic inter actions, and hydro gen bond formation.
Organotins cause cellular proteasome inhibition and cell death. To investigate whether the organotins could inhibit cellu lar protea somal activity in intact cells, we chose human breast cancer MDAMB231 cells as a work ing model. The cells were plated in a 96well plate and then treated with 10 µmol/L of each butyltin or phenyltin compound, SnCl 4 (an inorganic tin as a comparison), or solvent (as a control) for 4 hr and the proteasomal CTlike activity was measured. We found that TPT was most potent among all the organotins tested and caused 63% proteasome inhibition under the experimental condition, whereas DPT, MPT, and TePT inhibited the pro teasomal CTlike activity by 35%, 24%, and 11%, respectively ( Figure 3A). Therefore, the rank of potency to inhibit cellular protea some activity by organic phenyltins was: TPT > DPT > MPT > TePT, consistent with that of these phenyltins to inhibit the purified 20S proteasomal CTlike activity (compare Figures 3A and 1A).
To verify the cellular proteasomeinhibi tory ability of the organotins, MDAMB231 cells were grown on 100mm dishes and then treated under the same condition (10 µmol/L of each butyltin or phenyltin compound for 4 hr). After the treatment, cells were col lected and the cell extracts were prepared for analysis of proteasome inhibition by measur ing accumulation of ubiquitinated proteins. We detected accumulation of ubiquitinated proteins mainly after treatment with TPT or TBT ( Figure 3B). Other organotins had much less effect, whereas SnCl 4 failed to increase the level of ubiquitinated proteins, compared with the solvent control ( Figure 3B). We have reported a ubiquitinated form of IκBα protein with MW of about 56 kDa (Chen et al. 2007). A similar p56 band appeared after treatment of TPT or TBT, as detected by the specific antibody to IκBα ( Figure 3B, UbIκBα). In comparison, other organotins and SnCl 4 showed no effect on accumulating the ubiquitinated IκBα protein ( Figure 3B).
To investigate whether proteasomal inhibition by organotins is associated with cell death induction, we investigated both morphologic changes and PARP cleavage. Morphologic changes (shrinking, blebbing) were observed mainly in the MDAMB231 cells treated with TPT or TBT (data not shown, but see Figure 4C). Consistently, treatment with TPT or TBT caused the disappearance of the intact PARP protein (116 kDa), associated with production of a cleaved PARP fragment ( Figure 3B). In con trast, other organotins were weaker cell death inducers and generated mild morphologic changes (data not shown). Only DPT treat ment caused a low level of PARP cleavage; other organotins and inorganic SnCl 4 had no effect ( Figure 3B).
We and others have shown that, associ ated with apoptotic commitment, Bax protein (p21/Bax) could be cleaved by calpain, pro ducing a p18/Bax fragment that then forms a homodimer p36/Bax (Gao and Dou 2000;Wood and Newcomb 2000). Treatment of MDAMB231 cells with TPT and TBT also caused an increase in levels of p36/Bax, asso ciated with decreased levels of p21/p18/Bax ( Figure 3B). In contrast, other organotins had much less effect ( Figure 3B).
We then examined whether TPT and TBT could specifically inhibit the pro teasomal CTlike activity in intact cells. MDAMB231 cells were treated with TPT or TBT at indicated concentrations for 12 hr, followed by measuring the three pro teasomal activities in cell lysates prepared. At 2.5 µmol/L, TPT inhibited the CTlike, PGPHlike, and trypsinlike activities of the cellular proteasomes by 72%, 33%, and 3%, respectively ( Figure 3C). Similarly, TBT also preferentially inhibited the CTlike activity over two other activities of the cellular protea somes (data not shown). This result confirms that both TPT and TBT selectively inhibit proteasomal CTlike activity.
A 6hr treatment of MDAMB231 cells with TPT at 1.0, 2.5, and 5.0 µmol/L also caused a dosedependent inhibition (by 30%, 40%, and 58%, respectively) of the protea somal CTlike activity ( Figure 4A). Similarly, treatment for 6 hr with TBT at 1.0, 2.5, and 5.0 µmol/L inhibited 8%, 39%, and 46% of the cellular proteasomal CTlike activity, respectively ( Figure 4A). Consistent with that, we detected increased levels of polyubiquit inated proteins, IκBα, and the ubiquitinated form of IκBα in a dosedependent manner in the cells treated with TPT ( Figure 4B) and TBT (data not shown), compared with the cells treated with inorganic SnCl 4 or sol vent. For example, levels of unmodified IκBα protein were increased by TPT at 1 and 2.5 µmol/L, and the ubiquitinated IκBα protein was accumulated by TPT at 2.5 µmol/L and was further increased by TPT at 5 µmol/L, associated with a decreased level of unmodi fied IκBα protein ( Figure 4B). Furthermore, cell death also occurred in a dosedependent fashion in the cells treated with TPT or TBT, as supported by cellular morphologic changes ( Figure 4C) and PARP cleavage into p85 and p65 fragments ( Figure 4B). We also observed cleavage of p21/Bax into p18/Bax and accu mulation of p36/Bax in the cells treated with TPT ( Figure 4B). These results demonstrate that TPT and TBT can inhibit cellular protea somal CTlike activity, resulting in activation of celldeath-associated proteases. However, treatment with either TPT ( Figure 4D) or  TBT (data not shown) at all the tested con centrations did not cause DNA strand breaks, as evidenced by negativity of TUNEL assay, suggesting that organotininduced cell death does not involve DNA damage.

Kinetic studies of proteasome inhibition by TPT.
To determine whether proteasome inhi bition or cell death is induced first by organ otins, we treated MDAMB231 cells with 5 µmol/L of TPT for different time points, followed by measurement of proteasome inhi bition and cell death. The proteasome inhibi tion by TPT started as early as 30 min, as shown by 35% inhibition of the proteasomal CTlike activity and accumulation of ubiquit inated proteins ( Figure 5A,B). By 1 hr, protea some activity decreased by 40% ( Figure 5A) and levels of ubiquitinated proteins were fur ther increased ( Figure 5B). The ubiquitinated proteins accumulated until the last time point (16 hr, Figure 5B). Although levels of unmod ified IκBα protein increased between 1 and 2 hr, that of ubiquitinated IκBα protein accu mulated between 6 and 16 hr ( Figure 5B). Importantly, cell death was not observed before 2 hr treatment with TPT, as shown by lack of caspase3/7 activation ( Figure 5A), PARP cleavage ( Figure 5B), and cellular mor phologic changes (data not shown). Only between 2 hr and 6 hr treatment with TPT was caspase3/7 activity increased (by 3 to 4fold; Figure 5A) and strong PARP cleavage bands detected ( Figure 5B). Calpain activation was also found after 2 hr treatment, as shown by the increased levels of p65/PARP (Pink et al. 2000) and p18/Bax, as well as p36/Bax ( Figure 5B). These results suggest that protea some inhibition by TPT contributes to cell death induction.
Organotins inhibit proteasome activity in multiple human cell lines. One of the in vivo targets of organotins is the immune system. We therefore investigated whether organo tins can inhibit proteasome activity in human peripheral blood Jurkat T cells. Jurkat cells were treated with TBT and TPT at concen trations ranging from 0.01 to 2.5 µmol/L for 8 hr, followed by measurement of the pro teasomal CTlike activity, cell death, and TUNEL assays. TPT inhibited 14%, 33%, 58%, 79%, and 91% of proteasomal CTlike activity at concentrations of 0.01, 0.1, 0.5, 1, and 2.5 µmol/L, respectively, and TBT showed similar dosedependent inhibition ( Figure 6A). Again, SnCl 4 at 5 µM had no inhibitory effect ( Figure 6A). Proteasome inhibition in Jurkat cells treated with TPT and TBT was accom panied with dosedependent cell death: TPT induced 12% and 77% cell death at 0.01 and 2.5 µmol/L, respectively, whereas TBT induced 10% and 71% cell death at 0.01 and 2.5 µmol/L ( Figure 6B). Again, SnCl 4 at 5 µM had no such effect ( Figure 6A,B). TUNEL assay showed again that both TBT and TPT did not cause DNA strand breaks in Jurkat T cells ( Figure 6C and data not shown), confirm ing that organotininduced cell death is DNAdamage independent. Similar results were also found in human prostate cancer LNCaP cells and human normal, nontransformed YT natu ral killer cells (data not shown).
Direct binding of TPT to the cellular proteasome. To provide direct evidence for binding of TPT to cellular proteasome, we treated MDAMB231 cells with 10 µmol/L TPT or solvent for 2 hr, followed by prepar ing the proteasome immunoprecipitates using an antibody to the proteasomal α2 subunit. Aliquots of the prepared proteasome immu noprecipitates were used for measuring the associated proteasomal activity ( Figure 7A), β5 subunit protein ( Figure 7B), and the bound total Sn level ( Figure 7C). A lower level of CTlike activity was found in the protea some complexes prepared from the cells treated with TPT, compared with the con trol ( Figure 7A), confirming the strong pro teasomeinhibitory activity of TPT in cells. Equal protein levels of β5 subunit were found in the aliquots of the two proteasome immuno precipitate preparations ( Figure 7B), indicat ing that the decreased proteasome activity in TPTtreated sample preparation is not due to loss of the proteasome complex. Indeed, about 2.7 pM total Sn was found only in the aliquot of the proteasome complexes prepared from cells treated with TPT and not in the control ( Figure 7C). As a nonspecific binding control, no Sn was detected in aprotinin com plexes prepared from the above TPTtreated cells ( Figure 7C).
Because TPT is relatively stable and can be slowly degraded to DPT, MPT, and inorganic  tins only by human cytochrome P450 iso forms (Ohhira et al. 2006), and because only TPT at 10 µmol/L could cause 50% inhibi tion of proteasomal activity ( Figures 3A, 4A), we believe that most cellular Sn detected in the proteasome immunoprecipitates under our experimental condition (within 2 hr treat ment) should be TPT. Therefore, we conclude that the detected Sn represents TPT binding to the proteasome, which is responsible for the decreased proteasomal CTlike activity in the cells. These data strongly suggest that the proteasome is a direct cellular target of TPT.

Discussion
In the present study, we have provided several lines of evidence that suggest the proteasome as an important cellular target of environmen tal toxic organotins. We have shown that TPT binds the proteasomal CT β5 site by in silico docking analysis ( Figure 2) and that TPT potently and preferentially inhibits CTlike activity of purified 20S proteasomes ( Figure 1) and cellular 26S proteasomes (Figures 3-6) in an irreversible manner. We have also shown direct binding of TPT to the proteasome in cells (Figure 7). We found that phenyltins could inhibit both purified ( Figure 1A) and cellular pro teasomal CTlike activity ( Figure 3A). The rank of inhibition potency was TPT > DPT > MPT, TePT. The various proteasome inhibitory potencies of these phenyltins were roughly associated with their abilities to increase ubiquitinated proteins and the ubiq uitinated form of IκBα ( Figure 3B). Similar conclusions can be drawn for the butyltins.
It has been well documented that the interaction and modification of the OH group on Thr1 of proteasome β5 subunit are criti cal for the irreversible inhibition of CTlike activity (Fenteany et al. 1995). Results from our docking study suggest that Sn present in TPT might coordinately interact with the O γ atom of Thr1 of the proteasome β5 subunit (Figure 2). This hypothesis was supported by our experimental data that TPT irrevers ibly inhibited proteasomal CTlike activity (Figure 1). We therefore proposed a mecha nism for the interaction between TPT and proteasome β5 subunit in which the Sn atom in TPT could be attacked nucleophilically by Thr1 O γ , and the hybridization state of Sn could then change from sp3 to dzsp3, which allows the Sn to form a new coordinate bond with Thr1 O γ (Figure 2). During this pro cess, the positions of phenyl groups might need only a minor change from their original tetrahedral conformation ( Figure 2C).
The following arguments are consistent with the idea that TPTmediated protea some inhibition is functional and responsi ble for the observed cell death. First, when MDAMB231 cells were treated with TPT (or TBT), proteasome inhibition, calpain or caspase activation, and cell death induction were increased in both a dose and time dependent manner (Figures 3-5). Second, proteasome inhibition occurred before cell death ( Figure 5), as demonstrated by multiple assays. Finally, similar data were obtained in human peripheral blood Jurkat T cells and other cell lines (Figure 6 and data not shown).
The most significant finding of the pres ent study is the detection of Sn in the iso lated proteasome complexes from the cells treated with TPT (Figure 7). We found no inter action of Sn and proteasomes in pro teasome immunoprecipitates we prepared from the cells treated with the control solvent ( Figure 7C) or in aprotinin (a serine protease inhibitor) complexes we prepared from the cells treated with TPT ( Figure 7C), demon strating a specific interaction. Based on the proteasome inhibition potency of phenyltins and the known pharmacokinetics of TPT metallization, we conclude that the detected Sn in the isolated proteasome complexes rep resents the bound TPT. These data strongly suggest that TPT directly binds to the cellular proteasomal β5 subunit and inhibits CTlike activity in cells.
We found that TPT (and TBT) prefer ably inhibited CTlike activity of purified 20S proteasomes ( Figure 1C) and cellular protea somes ( Figure 3C), suggesting that β5 sub unit is a specific target for organotins. This finding is significant because inhibition of the proteasomal β5 CTlike activity is associated with cell apoptosis Lopes et al. 1997). Indeed, cell death occurred after TPT or TBT treatment. It appears that organotins induce at least some type of apoptosis because caspase3 activation and p85 PARP cleav age fragment were observed (Figures 4, 5). However, cell death seems to not involve DNA damage because TUNEL assays showed negative results (Figures 4, 6). Because calpain is associated with necrosis and is activated by TPT and TBT, we also suspected occurrence of necrosis under our experimental conditions. Taken together, we conclude that organo tins can induce necrosis and some caspase dependent, DNAdamage-independent cell death. Finally, because organotins have similar effect on LNCaP cells (containing wildtype p53 gene), MDAMB231 (mutant p53) cells, and Jurkat T cells (mutant p53), we suggest that organotins kill cells via a p53indepen dent pathway. This hypothesis is consistent with our previous report about proteasome inhibitors inducing tumor cell death in a p53 independent manner ).
Our present findings may also provide an explanation for some previous observations. TPT and TBT have been reported to induce apoptosis in various cell systems (Mundy and Freudenrich 2006;Stridh et al. 1999), and some potential targets of organotins include NFκB (Marinovich et al. 1996) and Bax (Zhu et al. 2007). Our results suggest that these previous observations might be down stream of proteasome inhibition because deg radation of both IκB and Bax is regulated by the proteasome pathway (Dou and Li 1999;Li and Dou 2000). Furthermore, exposure to TPT or TBT has been shown to induce imposex, a pseudohermaphroditic condition, in female sea animals (Smith 1996). Because cytochrome P450 aromatase is the critical enzyme to catalyze the conversion of andro gen to 17βestradiol in eukaryote cells, inhi bition of aromatase activity by TBT or TPT was often assumed to be responsible for the development of masculinized organs in female animals (Spooner et al. 1991). We suggest that inhibition of aromatase activity observed in organotinexposed animals is due to inhibi tion of the proteasomes because proteasome inhibition causes upregulation of the tran scription factors USF1/2 (upstream stimula tory factors 1 and 2), which in turn suppress transcription of the hCYP19/aromatase gene (Jiang and Mendelson 2005).  In summary, the results from our present study strongly suggest that the proteasome is one of the molecular targets of environmental toxic organotins in human cells. The connec tion of organotin exposure to cellular protea some inhibition provides a novel mechanism for environmental influences on human or animal health.